Rigorous results for timelike Liouville field theory

This paper establishes a rigorous mathematical framework for Gaussian random variables with negative variance to prove the timelike DOZZ formula for 3-point correlation functions and derive expressions for $k$-point functions in timelike Liouville field theory, demonstrating their convergence to correct semiclassical limits.

The Gist

Imagine you are trying to understand the fabric of the universe. In physics, there's a famous theory called Liouville Field Theory. Think of this theory as a recipe for how a 2D surface (like a crumpled piece of paper) wiggles and fluctuates.

For a long time, physicists had two versions of this recipe:

1. The "Spacelike" Version: This is the safe, well-behaved version. It's like baking a cake where the ingredients mix nicely. Mathematicians have already figured out exactly how to calculate the results of this recipe.
2. The "Timelike" Version: This is the wild, dangerous cousin. It's closer to the real theory of Quantum Gravity (how space and time behave at the tiniest scales). But there's a catch: in the math, the "energy" term has a negative sign.

The Problem: The "Negative Variance" Paradox

In everyday life, if you measure how much a random variable (like the height of people in a room) varies, the number is always positive. You can't have a "negative spread."

In the "Timelike" version of the theory, the math requires us to use a Gaussian random variable with negative variance.

• The Analogy: Imagine trying to bake a cake where the recipe says, "Add -2 cups of flour." If you try to do this with normal math, the batter explodes, or the numbers become imaginary and nonsensical. Physicists had been using "hand-waving" tricks to get answers, but no one had proven that these tricks actually worked or that the answers were real.

The Solution: A New Kind of Math

Sourav Chatterjee, a mathematician at Stanford, wrote this paper to fix that mess. He didn't just wave his hands; he built a brand new mathematical framework to handle these "negative variance" numbers.

Here is how he did it, using a simple metaphor:

The "Wrong Way" vs. The "Right Way"

• The Wrong Way (Naive Approach): Imagine you have a function (a machine) that takes a number $x$ and spits out a result. To handle the "negative" problem, you might try to plug in an imaginary number ($i$) directly into the machine.
  • The Result: The machine breaks. It spits out a result that is partly imaginary (like $3 + 4i$). But in physics, the final answer for a real-world measurement must be a real number (like just $5$). This approach failed.
• The Right Way (Chatterjee's Approach): Instead of plugging the imaginary number into the output, Chatterjee changed the input first. He took the function itself, turned it inside out (analytically continued it), and then plugged in the imaginary number.
  • The Result: The machine works perfectly! The imaginary parts cancel each other out, and you get a clean, real number. It's like realizing that to get a real shadow, you don't just shine a light on a ghost; you have to change the angle of the light source first.

The Main Achievements

With this new math tool in his belt, Chatterjee proved three major things:

1. The "DOZZ Formula" is Real
Physicists had a famous formula (called the DOZZ formula) that predicted how three points on this wiggly surface interact. For the "Spacelike" version, this was proven. For the "Timelike" version, they just guessed the formula by swapping some letters ($b$ to $ib$).

• The Proof: Chatterjee proved that this guess was actually correct, but only under specific conditions (like a "charge neutrality" rule, which is like ensuring the total weight of ingredients in a recipe balances out). He showed that the formula isn't just a lucky guess; it's a rigorous mathematical fact.

2. The "Semiclassical" Limit (The Big Picture)
In physics, there's a concept called the semiclassical limit. Imagine zooming out so far that the quantum "fuzziness" disappears, and the system behaves like a smooth, classical object (like a planet orbiting a star).

• The Discovery: Chatterjee showed that as you zoom out, the "Timelike" theory settles down into a specific shape. This shape isn't just random; it solves a specific equation that describes how gravity works in 2D.
• The Twist: The "smooth shape" the theory settles into isn't made of real numbers; it requires a complex number (a number with an imaginary part) to exist. However, when you calculate the actual physical metric (the shape of the universe), the imaginary parts cancel out, leaving a real, physical geometry. It's like a magician's trick where the "ghost" (the imaginary number) is necessary to make the "real" object appear.

3. Why This Matters
This paper is a bridge. It connects the messy, wild world of Quantum Gravity (where space and time are fuzzy and weird) with the clean, rigorous world of Mathematics.

• Before this, "Timelike" Liouville theory was like a ghost story: everyone talked about it, but no one could prove it existed.
• Now, Chatterjee has built a cage for the ghost, proving it exists, showing exactly what it looks like, and proving that it behaves exactly as the physicists predicted.

Summary in One Sentence

Sourav Chatterjee invented a new way to do math with "negative probabilities," proving that a wild, dangerous theory of quantum gravity actually works, produces real answers, and settles down into a predictable shape when you look at it from far away.

Technical Summary

1. Problem Statement and Context

Background:
Liouville field theory (LFT) is a fundamental two-dimensional conformal field theory (CFT) and a model for 2D quantum gravity. Standard (spacelike) LFT has been rigorously constructed using probability theory, specifically Gaussian Multiplicative Chaos (GMC), leading to proofs of the DOZZ formula (the 3-point correlation function) and connections to the Brownian map.

The Challenge:
"Timelike" Liouville field theory is a variant where the coupling constant $b$ is replaced by $ib$. This transformation reverses the sign of the kinetic term in the action.

• Physical Significance: The "wrong sign" kinetic term is a signature of models closer to 2D quantum gravity (e.g., JT gravity) than spacelike LFT.
• Mathematical Obstacle: The action becomes unbounded from below. In the path integral formulation, this corresponds to a "measure" with a density proportional to $e^{+\frac{1}{2}x^2}$ rather than $e^{-\frac{1}{2}x^2}$. This requires a theory of Gaussian random variables with negative variance ($N(0, -1)$), which does not exist in standard measure theory.
• Previous Attempts: Naive analytic continuation of the spacelike DOZZ formula (replacing $b \to ib$) fails to yield the correct timelike correlation functions. Previous rigorous work (Guillarmou et al.) addressed a compactified version of the theory, but the non-compact version remained open.

Objective:
To provide a rigorous mathematical construction of non-compact timelike Liouville field theory, prove the timelike DOZZ formula, derive $k$-point correlation functions, and establish the validity of the semiclassical limit.

---

2. Methodology

The paper develops a novel mathematical framework to handle "wrong sign" expectations and applies it to the path integral of timelike LFT.

A. Theory of Wrong Sign Gaussian Random Variables

The core innovation is the rigorous definition of expectations for random vectors $Z = (X_1, \dots, X_m, Y_1, \dots, Y_n)$ where $X_i \sim N(0, 1)$ and $Y_j \sim N(0, -1)$.

• Rejection of Naive Approaches: The author demonstrates that standard analytic continuation (computing $E[f(sY)]$ for $s>0$ and continuing to $s=i$) fails because it can yield complex values for real-valued functions, violating physical consistency.
• The Correct Definition: The expectation $E[f(Z)]$ is defined via analytic continuation of the function $f$, not the parameter.
  • Let $f: \mathbb{R}^{m+n} \to \mathbb{C}$. If $f$ admits an analytic continuation $\tilde{f}$ to a domain containing $\mathbb{R}^m \times (\mathbb{R} \cup i\mathbb{R})^n$, the expectation is defined as:
    $$E[f(Z)] := E[\tilde{f}(W_1, \dots, W_m, iW_{m+1}, \dots, iW_{m+n})]$$
    where $W_k \sim N(0, 1)$ are standard Gaussians.
• Key Properties: This definition ensures linearity, consistency with polynomials/exponentials, and crucially, that the expectation of a real-valued function remains real (proven via the Schwarz reflection principle).
• Applications: This theory is used to solve the backward heat equation and perform semiclassical approximations for "tilted" distributions.

B. Regularization and Path Integral Construction

To construct the correlation functions:

1. Regularization: The path integral is regularized by introducing a zero-mode cutoff ($\epsilon$), a spectral cutoff ($L$), and a damping factor ($\lambda$).
2. Expansion: The field $\phi$ is expanded in spherical harmonics. The path integral measure is transformed into a product measure over coefficients.
3. Wrong Sign Expectation: The regularized integral is expressed as an expectation with respect to the "wrong sign" Gaussian variables defined in Section 2.
4. Limits: The limits $\epsilon \to 0$, $L \to \infty$, and $\lambda \uparrow 1$ are taken carefully to derive explicit formulas.

C. Analytic Continuation of Structure Constants

The derived formulas are initially valid only in a "special region" of parameters (satisfying the charge neutrality condition). The paper uses the SL(2, C)-invariance of the theory and the properties of the special function $\Upsilon_b$ to analytically continue the structure constants to a larger domain, identifying non-trivial poles.

---

3. Key Contributions and Results

1. Rigorous Construction of Timelike LFT

The paper constructs the non-compact version of timelike Liouville field theory for parameters satisfying the charge neutrality condition:
$$w = \frac{Q - \sum_{j=1}^k \alpha_j}{b} \in \mathbb{Z}_{>0}$$
where $Q = b - 1/b$.

2. The Timelike DOZZ Formula (Theorem 1.3.2)

The author proves the timelike DOZZ formula for the 3-point correlation function $C(\alpha_1, \alpha_2, \alpha_3; z_1, z_2, z_3)$.

• The formula matches proposals by Schomerus, Zamolodchikov, and Kostov/Petkova.
• It is expressed in terms of the special function $\Upsilon_b$ and the Gamma function.
• Significance: This is the first rigorous proof of this formula for the non-compact theory, confirming that the "wrong sign" path integral yields the physically predicted structure constants.

3. General $k$-Point Correlation Functions (Theorem 1.3.1)

A general formula for $k$-point correlation functions ($k \ge 3$) is derived.

• The result takes the form of a Coulomb gas integral (similar to the spacelike case but with different exponents and a complex phase factor $e^{-i\pi w}$).
• The formula involves an integral over $w$ points on the sphere (or plane) with specific interaction potentials.

4. Semiclassical Limit with Heavy Operators (Theorem 1.3.3 & 1.3.5)

The paper analyzes the limit $b \to 0$ while scaling charges $\alpha_j = \hat{\alpha}_j/b$ and cosmological constant $\mu = \hat{\mu}/b^2$ (heavy operators).

• Variational Principle: The limit of the log-correlation function is shown to be determined by the minimization of a functional $S(\rho)$ involving entropy, interaction energy, and linear terms.
• Critical Points: The paper proves that the minimizer $\hat{\rho}$ corresponds to a complex-valued critical point $\hat{\psi}$ of the timelike action $J$.
• JT Gravity Connection: The critical point equation derived is shown to be the equation of motion for Jackiw-Teitelboim (JT) gravity with point charges.
  • Unlike spacelike LFT (which yields negative curvature), the timelike theory yields positive curvature, consistent with JT gravity requirements.
  • The critical point $\hat{\psi}$ is necessarily complex (specifically, it has a constant imaginary part of $i\pi/2$), explaining why the classical action has no real critical points in this regime.

5. Pole Structure

The paper identifies the poles of the 3-point function. It distinguishes between "trivial poles" (at the boundary of the parameter region) and "non-trivial poles" arising from the zeros of the $\Upsilon_b$ function in the denominator of the DOZZ formula.

---

4. Significance

• Mathematical Physics: This work bridges a major gap between the rigorous probabilistic construction of spacelike LFT and the physically motivated but mathematically elusive timelike LFT. It establishes a rigorous foundation for "wrong sign" Gaussian theories.
• Quantum Gravity: By proving that the timelike theory yields the correct semiclassical limit corresponding to JT gravity (with positive curvature), the paper validates timelike LFT as a viable toy model for 2D quantum gravity.
• New Mathematical Tools: The development of the theory of "wrong sign" Gaussian random variables provides a new toolset for handling path integrals with unbounded actions, potentially applicable to other models in quantum gravity and statistical mechanics.
• Resolution of Conjectures: The work confirms long-standing conjectures regarding the DOZZ formula and the semiclassical behavior of timelike LFT, resolving issues where naive analytic continuation failed.

In summary, Sourav Chatterjee provides a rigorous, non-perturbative construction of timelike Liouville field theory, proving that its correlation functions are well-defined, satisfy the expected DOZZ formula, and correctly reproduce the semiclassical physics of 2D quantum gravity.