Fate of Bouncing Singularities at Finite $G_N$

Technical Summary: Fate of Bouncing Singularities at Finite $G_N$

Problem Statement
In the semiclassical approximation of holographic Conformal Field Theories (CFTs) dual to Einstein-Hilbert gravity, thermal two-point functions of large-dimension scalar operators exhibit "bouncing singularities." These singularities arise from complexified geodesics that travel from the boundary, reflect off the black hole singularity, and reach the second asymptotic boundary. In the gravity approximation, these trajectories correspond to singularities in the analytically continued correlator at a specific complex time $\tau_1 = -\beta/2 + i(\beta/2)\cot(\pi/d)$.

The central question addressed is the status of these singularities in the exact theory at finite $N$ (finite central charge $C_T$, or finite Newton constant $G_N$). While the gravity approximation predicts a divergence at $\tau_1$, the exact CFT correlator must be bounded and analytic in the physical strip $0 < \text{Re}(\tau) < \beta$. The paper investigates whether the singularity persists on the "second sheet" of the analytically continued correlator when $G_N$ is finite, or if it is resolved by quantum gravity effects.

Methodology
The author employs rigorous analyticity properties of unitary CFTs to extend the domain of the thermal two-point function beyond the physical sheet. The analysis relies on two primary mathematical techniques:

1. Thermal Operator Product Expansion (OPE) and Summation by Parts:
   The thermal OPE is expressed as a sum over scaling dimensions $\Delta$: $H(r) = \sum a_\Delta (r/\beta)^\Delta$. While this series converges for real Euclidean separations $0 < r < \beta$, analytic continuation to complex time ($\tau = r e^{i\theta}$) introduces phases $e^{i\theta\Delta}$ that could potentially cause divergence if the coefficients $a_\Delta$ do not have definite signs.
   To prove convergence on the second sheet (where the bounce singularity would reside), the author utilizes a "summation by parts" trick. By comparing the series at the target complex point $\tau$ with a convergent series at a slightly larger real separation $r_0$ (where $r < r_0 < \beta$), the author constructs a weight factor $e^{-w\Delta}$ (with $\text{Re}(w) > 0$). This weight ensures exponential decay, proving that the analytically continued sum converges absolutely and defines an analytic function, provided $|\tau| < \beta$.

2. Partition Function Bounds via Path Integral Geometry:
   A second, independent bound is derived by cutting the thermal path integral along a sphere of diameter $r_0$ enclosing the operator insertions. This separates the system into an interior state $|u(r)\rangle$ (prepared by the operators in a ball) and an exterior state $\langle \Psi_{r_0}|$ (the thermal exterior).
   The correlator is the overlap $\langle \Psi_{r_0} | e^{i\theta D} | u(r) \rangle$, where $D$ is the dilatation operator. Using the Cauchy-Schwarz inequality, the magnitude of the correlator is bounded by the product of the norms of these two states. The norm of the exterior state is related to a partition function on a connected sum manifold (two thermal cylinders joined at a neck), while the interior norm is a vacuum four-point function.

Key Contributions and Results

• Absence of Singularity at Finite $N$: The primary result is that the bouncing singularity predicted by Einstein-Hilbert gravity is absent in any unitary CFT with finite $N$. The analytic continuation to the second sheet yields a finite, analytic function for $|\tau| < \beta$. Since the bounce location $\tau_1$ satisfies $|\tau_1| < \beta$ for dimensions $d=3, 4, 5$, the correlator is regular at this point in the exact theory.
• Failure of the Gravity Approximation: The paper demonstrates that the failure of the gravity approximation is not merely a feature of the second sheet but manifests on the physical sheet as well. The gravity approximation predicts an OPE radius of convergence of $\beta/\sqrt{2}$ (due to the stress-tensor sector), whereas the exact CFT requires convergence up to $\beta$. This implies that the exact OPE data (dimensions and coefficients) must deviate from the semiclassical gravity prediction at sufficiently high scaling dimensions, even before reaching the bounce point.
• Quantitative Bounds:
  • Partial Sum Bound: The author derives a bound on the correlator at $\tau_1$ in terms of the maximum partial sum of the OPE at $r_0$. By estimating how the exact OPE coefficients depart from the gravity approximation at a critical dimension $n^*$, the bound scales as an inverse power of $G_N$ (e.g., $|H(\tau_1)| \lesssim G_N^{-\alpha}$). This suggests the "singularity" is replaced by a large but finite peak.
  • Partition Function Bound: The Cauchy-Schwarz bound yields an estimate scaling as $\exp(N^2)$ (or $\exp(C_T)$). While mathematically rigorous, the author notes this bound is likely very loose and practically weak, as it does not account for cancellations between terms.

Significance and Claims
The paper claims to resolve the fate of bouncing singularities by establishing that they are artifacts of the semiclassical limit ($N \to \infty$) and do not exist in the exact finite-$N$ theory. The significance lies in:

1. Rigorous Analyticity: Proving that CFT correlators can be analytically continued to the second sheet without singularities within the thermal circle radius, ruling out simple "smoothing" or shifting of the singularity.
2. Constraints on Quantum Gravity: The results imply that any consistent quantum gravity theory dual to a CFT must modify the high-dimension OPE data to ensure the radius of convergence is $\beta$, not $\beta/\sqrt{2}$.
3. Nature of the Singularity: The paper remains agnostic on the mechanism of resolution. It does not determine whether the singularity is removed because the black hole singularity itself is resolved by quantum gravity, or because the back-reaction of infalling matter prevents the geodesic from reaching the singularity. The author notes that if no observable can reach the singularity due to back-reaction, the distinction between "no singularity" and "singularity hidden by back-reaction" may be semantic.

The work does not propose new experimental tests or specific future applications but serves as a theoretical consistency check on the holographic dictionary, highlighting the non-commutativity of the $N \to \infty$ limit and the OPE expansion.