🔢 mathematics

On the local conformal structure of Imaginary Liouville theory

This paper establishes Ward identities and Belavin-Polyakov-Zamolodchikov differential equations for Imaginary Liouville theory, providing a foundational step toward a rigorous conformal bootstrap program for non-rational conformal field theories with central charge less than one.

Original authors: Baptiste Cerclé, Romain Usciati

Published 2026-08-04

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Baptiste Cerclé, Romain Usciati

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 not as a collection of solid planets and stars, but as a giant, shimmering fabric that is constantly rippling, stretching, and twisting. In the world of theoretical physics, there is a special branch called "Conformal Field Theory" (CFT) that tries to describe how this fabric behaves when it's squished or stretched, but only in ways that preserve angles—like looking at a map where the shapes of countries stay the same even if the size of the continents changes. For decades, physicists have been obsessed with a specific, tricky version of this theory called "Liouville Theory." Think of it as the "Goldilocks" zone of these ripples: it's just complex enough to be interesting, but simple enough that we can actually solve the math equations for it. It's like having a perfect recipe for a cake that never fails.

However, there's a weird, shadowy cousin to this famous theory called "Imaginary Liouville Theory." The name sounds like it belongs in a sci-fi novel, but it's actually a serious mathematical puzzle. While the original theory describes a world with a certain amount of "energy" (or central charge) that is very high, this imaginary version describes a world with very low energy—so low that the usual rules of the game seem to break down. For a long time, scientists could only guess what this imaginary theory looked like by doing clever math tricks, but they couldn't prove it actually existed as a real, working system. It was like trying to build a house using only a blueprint that everyone agreed looked nice, but no one could actually lay the bricks.

This paper is the first time anyone has tried to actually lay those bricks. The authors, Baptiste Cerclé and Romain Usciati, take a bold new approach: instead of just guessing the answers, they build the theory from the ground up using a method called a "path integral." Imagine trying to predict the path of a drunk person walking home. You can't know exactly where they will step, but you can calculate the probability of them taking every possible route and adding them all up. That's what a path integral does for the universe's ripples. The authors set up this calculation for the "Imaginary" version, but they hit a snag: the math involves a mysterious "zero mode" (a kind of background hum of the universe) that needs to be integrated over a strange, winding path in the complex number system.

The main finding of this paper is that they successfully established that the theory has "local conformal invariance," provided they accept a specific conjecture about how fast a certain part of the math dies out at infinity. In plain English, this means that if you zoom in on a tiny patch of this imaginary fabric, the rules of the game look exactly the same as they do everywhere else. They derived two powerful mathematical tools called "Ward identities" and "BPZ equations" that act like a rulebook for how these ripples interact. These equations are the "Holy Grail" for physicists because they allow you to predict how any two particles will behave just by knowing how they behave individually. The authors showed that their path-integral construction obeys these rules, which is a massive step forward.

However, the paper is careful not to claim they have solved everything. They explicitly note that the Imaginary Liouville formula cannot be obtained from a conventional bosonic action, contrary to what some might expect. They also challenge a popular belief that this imaginary theory has a "purely continuous spectrum" (meaning it only has smooth, flowing energy levels). Instead, their new math suggests that, just like the original theory, there might be discrete, "stepped" energy levels hidden inside, which would change how we understand the theory's "spectrum."

The authors are very sure about their local rules (the Ward identities and BPZ equations) because they proved them using rigorous math, but only under the assumption of the conjecture mentioned earlier. They don't claim to have proven the entire global picture yet; they admit that proving the theory works on a global scale (the whole universe, not just a tiny patch) and proving that the "zero mode" integral definitely converges are still open problems. But by showing that the local rules work, they have provided the first solid evidence that the "conformal bootstrap" program—a grand plan to solve these theories by self-consistency—can actually work for these low-energy, imaginary worlds. It's like finding the first few solid stepping stones across a wide, foggy river; you haven't crossed the whole thing yet, but you now know the water isn't deep enough to drown you, and you can see the path forward.

Technical Summary: On the Local Conformal Structure of Imaginary Liouville Theory

Problem Statement
Imaginary Liouville theory has been proposed as a path integral construction for a non-rational conformal field theory (CFT) with central charge $c < 1$ , characterized by the "Imaginary DOZZ" formula for structure constants. While exact computations in specific regimes (such as charge neutrality) have been linked to Coulomb gas representations, a rigorous implementation of the conformal bootstrap program for this theory remained an open problem. Specifically, it was unclear whether the path integral definition, which involves integrating over a complex contour for the zero mode, possesses the intrinsic local conformal invariance required to derive Ward identities and Belavin-Polyakov-Zamolodchikov (BPZ) differential equations. Previous work established that the Imaginary DOZZ formula cannot be derived from a conventional bosonic action, necessitating a definition via steepest-descent cycles in the space of fields.

Methodology
The authors adopt a constructive probabilistic approach based on the path integral definition of Imaginary Liouville theory introduced in [UGRS26]. The core methodology involves:

Probabilistic Formulation: The theory is defined via a measure on the space of generalized functions on the Riemann sphere. The Liouville field is decomposed into a zero mode $c$ and a fluctuating part $\phi_0$ . The zero mode is integrated over a specific infinite oriented contour $U$ in the complex plane, while the fluctuating part is treated via the Gaussian Free Field (GFF).
Imaginary Gaussian Multiplicative Chaos (GMC): The interaction term in the action is interpreted as an instance of Imaginary GMC. Correlation functions are expressed as expectations involving the Laplace transform of this Imaginary GMC.
Assumptions and Conjectures: The derivation relies on Conjecture 1, which posits a specific decay rate for the Laplace transform of the Imaginary GMC as the zero mode shifts to $-i\infty$ . This conjecture ensures the convergence of the zero-mode integral. The authors also propose Conjecture 2, refining the Seiberg bounds (conditions on the vertex operator weights $\alpha_k$ ) required for the correlation functions to be well-defined.
Regularization and Renormalization: To handle singularities, the authors employ a two-step regularization:
- Smoothing the GFF at scale $\epsilon$ .
- Excluding neighborhoods of radius $\rho$ around vertex operator insertion points.
- Defining derivatives of the Liouville field and descendant fields via Wick products and Gaussian integration by parts.
Fusion Estimates (OPE): A critical technical component is the derivation of discrete Operator Product Expansion (OPE) estimates for Imaginary Liouville correlation functions. These estimates describe the asymptotic behavior when two insertion points collide, distinguishing the theory from ordinary Liouville theory due to the different behavior of Imaginary GMC.

Key Contributions and Results

Local Conformal Invariance: The paper establishes that Imaginary Liouville theory satisfies local conformal invariance. This is demonstrated by deriving the conformal Ward identities for descendant fields. The authors prove that the correlation functions satisfy:
$\left\langle L_{-n}^{\alpha_1} V_{\alpha_1}(x_1) \prod_{k=2}^N V_{\alpha_k}(x_k) \right\rangle = \sum_{k=2}^N \left( -\frac{\partial_{x_k}}{(x_k - x_1)^{n-1}} + \frac{\Delta_{\alpha_k}}{(x_k - x_1)^n} \right) \left\langle \prod_{k=1}^N V_{\alpha_k}(x_k) \right\rangle$
This result holds under the assumption of Conjecture 1 and specific bounds on the charges $\alpha_k$ .
Derivation of BPZ Equations: By specializing the Ward identities to degenerate fields (where $\alpha_1 \in \{-\beta/2, 2/\beta\}$ ), the authors derive the celebrated BPZ differential equations:
$\left( -\frac{1}{\alpha_1^2}\partial_{x_1}^2 + \sum_{k=2}^N \left( \frac{\partial_{x_k}}{x_1 - x_k} + \frac{\Delta_{\alpha_k}}{(x_1 - x_k)^2} \right) \right) \left\langle \prod_{k=1}^N V_{\alpha_k}(x_k) \right\rangle = 0$
This confirms that the path integral construction yields a theory with the expected differential structure for CFTs with central charge $c < 1$ .
Discrete OPE and Spectrum Implications: The authors prove a discrete OPE expansion (Lemma 3.4) for the product of two vertex operators. Unlike the continuous spectrum often assumed for Imaginary Liouville theory, these estimates suggest the presence of discrete contributions to the spectrum, analogous to ordinary Liouville theory. The expansion involves a finite sum of Virasoro descendants and a leading term with a shifted charge.
KPZ Identity and Zero-Mode Invariance: The proof of the Ward identities relies on a KPZ identity (Proposition 3.5), which relates the derivative of the correlation function with respect to the cosmological constant $\mu$ to the insertion of a vertex operator. Crucially, the authors show that local conformal invariance arises from the invariance of the zero-mode integral under shifts in the imaginary direction, whereas global conformal covariance would require invariance under real shifts (which is not satisfied by the chosen contour).

Significance and Claims
The paper claims to provide the first analytical evidence (both in physics and mathematics) that the conformal bootstrap program can be rigorously carried out for CFTs with central charge less than one. It represents the first investigation into the intrinsic properties of Imaginary Liouville theory beyond exact computations in special regimes.

The results serve as a foundational step toward a rigorous derivation of the Imaginary DOZZ formula within the path-integral framework. By establishing the Ward identities and BPZ equations, the authors demonstrate that the path integral construction is consistent with the axioms of local conformal symmetry.

Limitations and Open Problems
The authors remain modest regarding the completeness of their work. They explicitly state that:

Global conformal covariance (or Weyl invariance) of the correlation functions has not been proven; this would require invariance of the zero-mode integral under real shifts, which is not guaranteed by the current contour choice.
The convergence of the zero-mode integral relies on Conjecture 1, which is supported by heuristic and numerical evidence but not rigorously proven within this work.
The full bootstrap formula expressing arbitrary correlation functions in terms of conformal blocks is not yet derived, though the BPZ equations provide a route toward it.

In summary, the work successfully bridges the gap between the probabilistic path integral definition and the algebraic structures of CFT (Ward identities, BPZ equations) for Imaginary Liouville theory, contingent upon the decay properties of Imaginary Gaussian multiplicative chaos.

Technical Summary: On the Local Conformal Structure of Imaginary Liouville Theory

More like this