⚛️ high-energy theory

Probabilistic construction of non compactified imaginary Liouville field theory

This paper proposes a probabilistic construction of non-compactified imaginary Liouville Field Theory based on a real Gaussian Free Field, demonstrating through exact results and numerical simulations that it reproduces the imaginary DOZZ structure constants without requiring a neutrality constraint.

Original authors: Romain Usciati, Colin Guillarmou, Remi Rhodes, Raoul Santachiara

Published 2026-08-05

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Romain Usciati, Colin Guillarmou, Remi Rhodes, Raoul Santachiara

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 the universe as a giant, bouncy trampoline. In physics, we often try to describe how this trampoline ripples and warps using math. Sometimes, these ripples behave like real, solid waves you can touch; other times, they behave like "imaginary" waves—mathematical ghosts that don't exist in our physical world but are incredibly useful for solving puzzles about how the universe might work. This paper lives in the world of Quantum Field Theory, a branch of physics that tries to explain how tiny particles and forces interact. Specifically, it tackles a tricky problem called Liouville Field Theory. Think of this theory as a set of rules for how a surface (like our trampoline) bends and twists. For a long time, scientists knew how to write these rules when the surface was "real," but when they tried to use "imaginary" numbers to describe certain exotic situations—like the behavior of loops in magnetic materials or the edges of strings in string theory—the math broke down. The old rules required a strict "neutrality" rule, like a scale that had to be perfectly balanced with equal weights on both sides, or the whole calculation would collapse. This made it impossible to describe many interesting, unbalanced scenarios.

Now, enter the team of researchers who decided to rebuild the math from the ground up. They proposed a new way to construct these "imaginary" theories using a tool called Gaussian Free Fields, which is essentially a fancy name for a random, wiggly surface that pops up naturally in probability. Their big idea? Instead of forcing the surface to be balanced (neutral), they let it wander freely, but they changed how they added up the numbers. They used a special, winding path in the complex number system (a bit like taking a detour through a parallel universe to avoid a traffic jam) to integrate the equations. The result is a theory that doesn't need that strict balancing act. They tested this new theory by running massive computer simulations on a sphere and comparing the results to a famous, theoretical formula known as the "imaginary DOZZ structure constants." The simulations matched the theory almost perfectly, suggesting that this new, unbalanced approach is the correct way to describe these imaginary worlds.

The Story of the Wobbly Trampoline

Let's dive into the details. For the past decade, mathematicians and physicists have been building a rigorous, step-by-step definition of how these quantum surfaces behave. They've done a great job with the "real" versions, where the math is stable and predictable. But when they tried to stretch these rules to the "imaginary" side (where a parameter called $b$ becomes a pure imaginary number, $i\beta$ ), things got messy. The old method, known as Compactified Imaginary Liouville Theory, was like a game with strict rules: you could only play if the total "charge" of your particles added up to zero. If you tried to look at a system that wasn't balanced, the math said the answer was zero. This was a dead end for many real-world problems, like understanding how different parts of a magnetic material connect or how strings vibrate in higher dimensions.

The authors of this paper, Romain Usciati, Colin Guillarmou, Remi Rhodes, and Raoul Santachiara, asked a bold question: What if we don't force the system to be neutral? They proposed a new construction based on a real, non-compactified Gaussian Free Field. In plain English, they stopped pretending the surface was stuck on a circle and let it roam freely on a flat plane. But here's the catch: if you just let it roam, the math explodes into infinity. To fix this, they introduced a clever trick involving the path of integration.

Imagine you are trying to cross a river to get to a treasure chest. The direct path (the standard way of doing math) leads to a waterfall where the numbers blow up. The authors found a secret bridge—a specific, winding path in the complex number plane shaped like a "U" (or a Hankel contour). By walking along this U-shaped bridge, the numbers stay calm and finite. This allows them to calculate the three-point functions of the theory. Think of a three-point function as a measure of how likely it is for three specific events to happen together on this wobbly surface.

The paper shows that when they use this U-shaped path, the results they get match the imaginary DOZZ structure constants. These constants are like the "recipe cards" for the universe's most exotic interactions. They had been derived before using a different method called the "bootstrap approach," which is like solving a puzzle by guessing the shape of the pieces based on how they fit together, without actually building the pieces. The authors' work is exciting because it provides the first explicit, "Lagrangian" (recipe-based) construction that reproduces these constants without needing the strict neutrality rule.

The Proof is in the Pudding (and the Simulations)

How do we know this new path works? The team didn't just wave their hands; they did the heavy lifting. First, they looked at the simplest case: a circle. They used exact mathematical results to prove that their method works perfectly there, showing that the "one-point function" (the behavior of a single point on the circle) follows the right rules. They even showed that their new theory connects smoothly to the old, real Liouville theory, acting like a bridge between two different worlds.

Then, they moved to the big test: the sphere. Calculating the three-point function on a sphere is much harder because the surface is two-dimensional and the math gets incredibly complex. To tackle this, they ran numerical simulations. They generated thousands of random "wiggly surfaces" (using a method called Gaussian Multiplicative Chaos) and calculated the averages. They compared their simulation results directly to the imaginary DOZZ formula.

The results were striking. The paper reports that their numerical data "agrees remarkably well" with the theoretical formula across several orders of magnitude. In the graphs, the dots from their computer simulations line up almost perfectly with the red curve of the theoretical prediction. There are some limits, of course. When the numbers get very close to a "pole" (a point where the value shoots up to infinity), the simulations get a bit shaky. The authors explain this is likely because the computer struggles to sample the extremely rare, wild events that cause those spikes. But in the vast, safe territory away from those spikes, the match is excellent.

What This Means for the Future

This paper doesn't claim to have solved every mystery in physics. It doesn't say, "We have found the final theory of everything." Instead, it offers a solid, probabilistic construction of a theory that was previously only a ghost. It proves that you can define these imaginary fields without the restrictive neutrality condition, opening the door to studying observables in geometric statistical models and two-dimensional quantum gravity that were previously out of reach.

The authors are careful to note that while their three-point functions match the bootstrap solution, there is still a big question mark hanging over the four-point function. Does this new theory break down when you add a fourth point? The paper leaves this as an open question for future investigation. But for now, they have successfully built a sturdy bridge across the river of imaginary numbers, showing us that the "unbalanced" world of imaginary Liouville theory is not only real but also beautifully consistent with the rules of the universe.

Technical Summary: Probabilistic Construction of Non-Compactified Imaginary Liouville Field Theory

Problem Statement
The paper addresses a long-standing gap in the probabilistic construction of two-dimensional Conformal Field Theories (CFTs). While the real Liouville Field Theory (LFT) with parameter $b \in \mathbb{R}$ is rigorously defined via Gaussian Free Fields (GFF) and satisfies the bootstrap equations, the regime where the parameter is purely imaginary ( $b = i\beta, \beta \in \mathbb{R}$ ) has remained elusive. Previous attempts to define "imaginary LFT" relied on a compactified version (CILT), where the field takes values on a circle. This compactification imposes a strict "neutrality condition" (selection rules) on correlation functions, rendering the theory insufficient for describing observables in geometric statistical models (e.g., O(n) loop models, Potts models) or applications in 2D quantum gravity where such constraints are absent. The central question is whether a probabilistic construction of non-compactified imaginary LFT exists that reproduces the "imaginary DOZZ" structure constants without enforcing neutrality.

Methodology
The authors propose a probabilistic construction of imaginary LFT ( $LFT_{i\beta}$ ) based on a real, non-compactified Gaussian Free Field. The core of their approach involves:

Action and Field Decomposition: They define the action $S^{im}_L$ on a surface $\Sigma$ with metric $g$ , featuring a bulk cosmological constant $\mu$ and boundary constant $\mu_B$ . The field $\phi$ is decomposed into a GFF $X_g$ and a constant zero mode $c$ . Crucially, unlike standard constructions where $c$ is real, the authors allow $c$ to take general complex values ( $c \in \mathbb{C}$ ).
Imaginary Gaussian Multiplicative Chaos (GMC): The theory relies on the imaginary GMC, defined as the limit of regularized exponentials of the GFF. For the circle case, this is $M^S_\beta$ . The authors restrict $\beta$ to $(0, \sqrt{2})$ to ensure the existence of finite $L^2$ moments and exponential moments for the chaos variable.
Contour Integration (Hankel-type): The path integral over the zero mode $c$ $c$ is defined by integrating along a specific U-shaped (Hankel-type) contour $\mathcal{C} = U$ $C = U$ in the complex plane. This contour consists of the lines $[0, -i\infty)$ $[0, - i \infty)$ , $[0, 4\pi/\beta]$ $[0, 4 π / β]$ , and $[4\pi/\beta, 4\pi/\beta - i\infty)$ $[4 π / β, 4 π / β - i \infty)$ .
- This choice is motivated by Morse theory and the need to perform analytic continuation from the real LFT regime.
- The convergence of the integral relies on the asymptotic behavior of the moment generating function (Laplace transform) of the imaginary GMC, $G_\beta(\mu) = \mathbb{E}[e^{-\mu M_\beta}]$ , which decays as $\mu^{-4/\beta^2}$ for large $\mu$ .
Correlation Functions:
- 1-point function: Calculated on the unit disk with $\mu=0$ . The integral over the contour $U$ yields a closed-form expression involving Gamma functions.
- 3-point function: Calculated on the Riemann sphere ( $S^2$ ). The structure constants $C_\alpha$ are defined by integrating the generating function of the imaginary GMC with charge insertions over the same U-contour.

Key Contributions and Results

First Explicit Lagrangian Construction: The paper presents the first explicit Lagrangian field theory construction of non-compactified imaginary LFT that generates the imaginary DOZZ structure constants without requiring a neutrality constraint.
Analytic Continuation and Shift Relations:
- The authors demonstrate that the 1-point function of their theory satisfies the same shift relations as the real LFT, derived from the insertion of degenerate states.
- They establish a precise relation between the 1-point function of the standard real LFT (analytically continued to $b=i\beta$ ) and their new theory, mediated by a factor involving $\sin(\pi s)$ .
- When the parameter $s$ (related to the sum of charges) is an integer, the theory reduces to the Coulomb Gas (CG) integrals of the compactified theory, showing consistency with known results.
Agreement with Imaginary DOZZ Constants:
- The primary result is the definition of the 3-point structure constants via the path integral (Eq. 15).
- The authors conjecture that these constants coincide with the imaginary DOZZ structure constants ( $C^{ImDOZZ}_\alpha$ ), which are the unique solution to the bootstrap shift equations for central charge $c \le 1$ .
- Numerical Verification: Extensive numerical simulations were performed on the sphere. The results show "excellent agreement" between the numerically computed 3-point functions and the imaginary DOZZ formula across several orders of magnitude. The relative deviation is minimal except near the poles of the DOZZ function, where the authors attribute discrepancies to the difficulty of sampling the rare events (tails of the GMC distribution) that determine the pole positions.
Generalization: The framework is proposed for general surfaces $(\Sigma, g)$ , subject to Seiberg-type bounds ( $\alpha_j > Q$ and $s < \beta \lambda$ ) ensuring the convergence of the Laplace transform.

Significance and Claims
The paper claims to provide a rigorous probabilistic definition of the imaginary DOZZ structure constants, which have previously been defined only via bootstrap equations or as analytic continuations of Coulomb gas integrals. By constructing the theory directly from a non-compactified GFF, the authors offer a natural extension of the compactified imaginary Liouville theory (CILT).

The significance lies in:

Removing Neutrality Constraints: The theory describes observables in geometric statistical models (like Potts model connectivities) that do not satisfy the neutrality condition required by CILT.
Bridging Probabilistic and Bootstrap Approaches: It connects the rigorous probabilistic construction (GFF and GMC) with the analytic bootstrap solutions for $c \le 1$ .
Foundation for Future Work: The authors note that while the 3-point function is solved, the factorization properties of the 4-point function remain an open question, particularly regarding whether this non-compact bootstrap solution is an extension of Coulomb gas integrals (which the known non-compact solution in [RS15] is not).

The authors remain modest regarding the numerical limitations near the poles of the structure constants, attributing the loss of accuracy to the statistical difficulty of sampling the tails of the imaginary GMC distribution, rather than a failure of the theoretical definition.

The Story of the Wobbly Trampoline

The Proof is in the Pudding (and the Simulations)

What This Means for the Future

Technical Summary: Probabilistic Construction of Non-Compactified Imaginary Liouville Field Theory

More like this