Sampling the Graviton Pole and Deprojecting the Swampland

This paper introduces a primal bootstrap framework using finite-resolution sampling to derive improved projective bounds and new non-projective constraints on effective field theories with graviton poles, demonstrating that in five dimensions the EFT cutoff cannot exceed the Planck scale and revealing distinct extremal spectral structures characterized by quadratic Regge-like trajectories or sharp bands.

The Gist

Imagine you are trying to figure out the rules of a game by only watching the players from a distance. You can't see the players' faces or hear their whispers, but you can see how they move, where they stop, and how they bounce off each other. In the world of physics, this "game" is the universe, and the "players" are particles.

Physicists use a tool called Effective Field Theory (EFT) to describe these particles. Think of EFT as a rulebook that lists all the possible moves particles can make at low energies. However, this rulebook has "wildcard" cards called Wilson coefficients. These are numbers that tell us how strong the interactions are, but the rulebook doesn't tell us what those numbers should be.

The big question is: Which numbers are allowed? Are there rules that say, "You can't have a number this big," or "You can't have a negative number here"?

The Problem: The Gravity Glitch

For a long time, physicists had a great way to find these rules using a method called dispersion relations. It's like listening to the echo of a sound to figure out the shape of a room. By analyzing how particles scatter (bounce off each other), they could prove that certain numbers must be positive.

But then, they tried to include gravity.

Gravity is tricky because it's carried by a particle called the graviton. The graviton is massless and travels forever, creating a "singularity" (a mathematical infinity) in the equations when particles try to bounce directly back and forth. It's like trying to measure the echo in a room where the walls are made of infinite mirrors; the math breaks down.

To fix this, previous researchers used a technique called "smearing." Imagine trying to take a photo of a fast-moving, blurry object. Instead of trying to focus on one sharp point (which is impossible because of the blur), you take a slightly fuzzy, averaged picture. This "smearing" smoothed out the gravity problem, but it had a downside: it blurred the details. It could tell you the ratio of two numbers, but it couldn't tell you the absolute size of the numbers. It was like knowing a car is twice as fast as a bike, but not knowing if the car is going 20 mph or 200 mph.

The New Solution: Sampling the Grid

The authors of this paper, Guangzhuo Peng and colleagues, decided to try a different approach. Instead of "smearing" the data, they used "sampling."

Imagine you are trying to map a mountain range.

• The Old Way (Smearing): You take a wide, blurry satellite photo. You can see the general shape, but you can't see the exact height of the peaks.
• The New Way (Sampling): You send out a team of hikers to measure the height at specific, carefully chosen points. You don't measure every point (that would take forever), but you measure enough points to build a perfect 3D model.

They call this the Primal Bootstrap. They set up a grid of "sampling points" across the mathematical landscape. At each point, they check if the rules of physics (like energy conservation and causality) hold up.

The Catch: The mountain is steep near the "gravity peak" (the singularity). If you pick your hikers randomly, they might fall off the cliff. The authors realized they needed to be smart about where they placed their hikers. They used a special pattern (Chebyshev nodes) that puts more hikers near the dangerous edges where the math gets tricky, ensuring they don't miss anything important.

The Big Discovery: The "Swampland" Fence

Once they got their sampling method working, they found something amazing.

In the past, the "smearing" method suggested that you could have a theory where the energy scale of your physics (the "cutoff") was infinitely larger than the scale of gravity (the Planck scale). It was like saying you could build a skyscraper on a foundation that was infinitely small, as long as you didn't look too closely.

But the new sampling method showed that this is impossible.

They found a hard limit. In 5-dimensional space, they proved that the energy scale of your theory cannot be much larger than the Planck scale.

• The Analogy: Imagine trying to build a house of cards. You can build a few layers, but if you try to build a skyscraper, the cards will collapse. The authors found the exact height of the "card tower" before it collapses.
• The Result: They found that the ratio of the theory's energy scale to the Planck scale must be less than about 7.8. You can't just make the theory arbitrarily large; gravity forces it to stay small. This is a "Swampland" result—it tells us that theories that try to be too big belong in the "Swampland" (a place of inconsistent, impossible theories) rather than the "Landscape" (the set of real, possible universes).

The Surprise: Quadratic Trajectories

When they looked at the "extremal" solutions (the most extreme cases where the rules are pushed to the limit), they found a surprising pattern in the data.

Usually, physicists expect particles to line up in linear patterns (like a straight line on a graph), similar to how strings vibrate in string theory.

• The Old Expectation: A straight line.
• The New Reality: The particles lined up in quadratic curves (like a parabola, a "U" shape).

It's as if, instead of a straight road, the particles are driving on a curved ramp. Even more strangely, the shape of these curves followed a specific mathematical pattern that got tighter and tighter as you went up the ladder of particle types. This pattern emerged naturally from their math, without them forcing it to happen. It suggests that the universe has a hidden, curved structure that we haven't seen before.

Summary

1. The Problem: Gravity makes standard math tools break down, and previous fixes were too blurry to see the full picture.
2. The Fix: The authors developed a "sampling" method that checks the rules of physics at specific, smartly chosen points, avoiding the mathematical cliffs.
3. The Result: They proved that you cannot have a theory of gravity that is infinitely larger than the Planck scale. There is a hard ceiling.
4. The Surprise: The "extreme" versions of these theories organize themselves into curved, quadratic patterns, not the straight lines we usually expect.

In short, they built a better ruler to measure the universe, and it turned out the universe is much more constrained—and more curvy—than we thought.

Technical Summary

1. Problem Statement

The paper addresses the challenge of deriving rigorous bounds on Effective Field Theories (EFTs) coupled to dynamical gravity using the S-matrix bootstrap program.

• The Graviton Pole Obstruction: Standard dispersive positivity bounds rely on forward scattering amplitudes. However, in the presence of gravity, the $t$-channel graviton exchange introduces an infrared singularity ($1/t$) at $t=0$. This renders standard fixed-$t$ dispersion relations ill-defined without regularization.
• Limitations of Smearing: Previous work (e.g., Ref. [21]) resolved this by "smearing" the amplitude over wavepackets in impact-parameter space. While conceptually sound, smearing introduces non-local correlations that obscure the linear structure of linearized unitarity ($0 \le \rho_J(s) \le 2$). This makes it difficult to impose pointwise unitarity constraints and prevents the derivation of absolute (non-projective) bounds on Wilson coefficients, as the gravitational coupling cannot be scaled away.
• The Goal: Develop a framework that handles the graviton pole directly, allows for the imposition of linearized unitarity, and yields absolute bounds on EFT couplings (deprojecting the Swampland) while providing access to the extremal spectra.

2. Methodology: Primal Bootstrap via Finite-Resolution Sampling

The authors propose an alternative to smearing: a primal bootstrap framework based on finite-resolution sampling.

• Core Concept: Instead of averaging the amplitude over a region (smearing), the dispersive relations are treated as functional equalities sampled at a finite set of kinematic points (Chebyshev nodes) within the valid domain ($t \in (-1, 0)$ or crossing-invariant variable $a \in (-1/3, 0)$).
• Discretization:
  • The continuous spectral integral over mass ($s'$) is discretized into $N_\mu$ points.
  • The partial-wave expansion is truncated at a maximum spin $J_{\text{max}}$.
  • The spectral densities $\rho_J(s')$ are treated as optimization variables subject to linear constraints: $0 \le \rho \le 2$ (linearized unitarity).
• Formulation: The problem is reduced to a finite linear program (LP). The dispersive sum rules (relating IR EFT coefficients to UV spectral integrals) become linear equality constraints on the discretized spectral densities.
• Numerical Stability Strategy:
  • Chebyshev Sampling: Points are clustered near the singular endpoints to resolve rapid functional variations while avoiding the exact singularity at $t=0$ or $a=0$.
  • Correlated Truncation: A critical finding is that stability requires a specific correlation between the number of sampling points ($N_t$ or $N_a$) and the spin cutoff ($J_{\text{max}}$). Specifically, for fixed-$a$, stability is found in the window $2N_a \lesssim J_{\text{max}} \lesssim 3N_a$. If $J_{\text{max}}$ is too low, the graviton pole cannot be resolved; if too high, the system becomes underconstrained.
• Framework Choice: The authors utilize fixed-$a$ crossing-symmetric dispersion relations (Refs. [76–78]) rather than fixed-$t$ relations. Fixed-$a$ relations are numerically more stable, converge faster, and naturally incorporate crossing symmetry without needing complex "improved sum rules" to cancel lower-order terms.

3. Key Contributions

1. Resolution of the Graviton Pole without Smearing: Demonstrated that direct sampling of dispersive relations is a viable and robust alternative to smearing, provided numerical truncations are carefully correlated.
2. Direct Access to Extremal Spectra: Unlike dual methods or smearing approaches where spectral extraction is indirect, the primal LP formulation explicitly outputs the extremal spectral density $\rho_J(\mu)$ as part of the solution.
3. Deprojection of the Swampland: By imposing the upper bound of linearized unitarity ($\rho \le 2$), the authors break the homogeneity of the constraints. This allows for the derivation of absolute (non-projective) bounds on the EFT couplings, fixing their overall scale relative to the Planck mass, rather than just ratios.
4. Discovery of Quadratic Regge Trajectories: The method revealed a novel structure in the extremal spectra: instead of the linear Regge trajectories typical of string theory, the extremal solutions organize into quadratic Regge-like bands.

4. Key Results

A. Projective Bounds (Positivity Only)

• In dimensions $D \ge 6$, the sampling method reproduces known projective bounds obtained via smearing, validating the approach.
• In $D=5$, the method yields a slightly stronger bound on the ratio of Wilson coefficients ($g_2/8\pi G \gtrsim -16.5$) compared to the smeared result ($-18$).
• Extremal Spectra (Projective): The spectra populate the full allowed region but are dominated by states forming quadratic Regge-like trajectories in the $(J, \mu)$ plane, with strong suppression elsewhere. This contrasts with previous dual-bootstrap results which suggested isolated states in a triangular region.

B. Non-Projective Bounds (With Linearized Unitarity)

• Absolute Bound on Gravity: In $D=5$, imposing linearized unitarity leads to a finite upper bound on the dimensionless gravitational coupling:
  $$\frac{M}{M_P} \lesssim 7.8$$
  This implies the EFT cutoff scale $M$ cannot be parametrically separated from the Planck scale $M_P$. As $M \to M_P$, the dimensionless coupling becomes $O(1)$, and the EFT expansion breaks down.
• Extremal Spectra (Non-Projective): The extremal solutions exhibit a sharp, band-like structure.
  • The support of the spectral density forms narrow bands bounded by quadratic trajectories: $\mu_n \sim b_{0,n} + b_{1,n}J + b_{2,n}J^2$.
  • The leading coefficient $b_{2,n}$ follows an inverse-quadratic dependence on the band number $n$: $b_{2,n} \sim A/(B+n)^2$.
  • The spectral density is "sharp," taking values close to either 0 or 2 (the unitarity bound).

C. Numerical Stability and Loop Corrections

• Stability: The authors provide extensive convergence tests showing that bounds stabilize within the correlated window of $N_a$ and $J_{\text{max}}$.
• Loop Corrections: Preliminary analysis of one-loop graviton and EFT corrections suggests that while the shape of the allowed region deforms, the upper bound on the gravitational coupling ($M/M_P$) remains robust and largely unaffected by these corrections in the weak-coupling regime.

5. Significance

• Theoretical Insight: The emergence of quadratic trajectories is a surprising result. It suggests that the constraints of analyticity, crossing symmetry, and unitarity in gravitational EFTs dynamically select a spectrum distinct from the linear Regge trajectories of string theory, without assuming any specific UV completion.
• Swampland Constraints: The result $M/M_P \lesssim 7.8$ provides a concrete, model-independent constraint on the "Swampland." It suggests that EFTs coupled to gravity cannot have a parametrically large separation between the cutoff and the Planck scale, reinforcing the idea that gravity imposes sharp limits on effective theories.
• Methodological Advancement: The paper establishes primal sampling as a powerful tool for gravitational bootstrap problems, overcoming the technical hurdles of the graviton pole and enabling the study of non-projective bounds and extremal spectra that were previously inaccessible.

In summary, the paper successfully develops a robust numerical framework to "deproject" the Swampland, proving that gravity imposes absolute limits on EFT scales and revealing a unique, quadratic organization of extremal spectra.