⚛️ high-energy theory

The origin of amplitude chaos and strings

This paper analytically and numerically demonstrates that smooth, nondegenerate amplitude ensembles, including tree-level highly excited string scattering amplitudes, exhibit a robust quartic nearest-neighbor peak repulsion ( $p_{\rm nn}(S)\sim S^4$ ) that defines a local universality class dependent on the differentiability of the underlying functions rather than being unique to string theory.

Original authors: Pallab Basu, Anurag Sarkar

Published 2026-10-12

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Pallab Basu, Anurag Sarkar

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

In the quantum world, the behavior of particles is often described by a mathematical object called an amplitude. Think of this amplitude not as a single number, but as a complex, wavy landscape that changes depending on how particles collide and scatter. When physicists study highly excited strings—vibrating one-dimensional objects that are candidates for the fundamental building blocks of the universe—they find that these landscapes are filled with sharp peaks. These peaks represent the most likely outcomes of a collision. For decades, researchers have noticed that these peaks seem to avoid each other; they rarely sit right next to one another. This phenomenon, known as peak repulsion, has long been suspected to be a sign of chaos, similar to how the energy levels of a chaotic system repel one another. However, the reason for this avoidance in string theory has remained a mystery, largely because the standard mathematical tools used to explain chaos in other systems do not seem to apply here. The peaks are not random numbers drawn from a hat; they are the precise, calculated results of a specific physical formula.

A team of researchers at the University of the Witwatersrand in South Africa has now peeled back the layers of this mystery, revealing that the repulsion is not a random accident but a direct consequence of how smooth and well-behaved the mathematical functions describing these strings are. By analyzing the exact formulas for string collisions, they discovered that the probability of two peaks being very close to each other drops off incredibly fast as the distance between them shrinks. Specifically, if you halve the distance between two peaks, the chance of finding them that close drops by a factor of sixteen. This "quartic" relationship, where the probability scales with the fourth power of the distance, turns out to be a universal rule. It is not unique to string theory or even to the specific way strings vibrate. Instead, the researchers found that any smooth, well-defined mathematical landscape, whether it comes from string theory or a simple made-up model, will naturally produce this same four-power repulsion, provided the landscape is smooth enough.

The researchers arrived at this conclusion by treating the string amplitudes as a collection of functions and asking a simple question: what must happen locally for two peaks to squeeze together? They found that for two peaks to approach each other, the function describing the collision must pass through a very specific, constrained configuration. It is as if the landscape is forced to flatten out in a particular way to allow the peaks to get close. Because the functions involved are smooth and differentiable, the probability of this specific configuration occurring becomes vanishingly small as the peaks get closer. The team verified this idea using massive computer simulations of string collisions, checking thousands of different scenarios. They confirmed that the repulsion follows the predicted fourth-power law with high precision. Furthermore, they tested their theory on other types of mathematical landscapes, including random waves and models based on integer partitions, and found the same rule held true. This suggests that the "repulsion" is a fundamental property of smoothness itself, rather than a secret code hidden within the complex machinery of string theory.

The study also explored what happens when the smoothness of the landscape is broken. The researchers introduced a "cusp," a sharp corner or kink in the mathematical function, to see how the repulsion would change. They found that when the function is no longer perfectly smooth, the repulsion weakens. The probability of peaks being close together no longer drops off as steeply as the fourth power; instead, it follows a gentler curve that depends on how sharp the kink is. This finding is crucial because it defines the boundary of the rule. The quartic repulsion is not a magic property of all mathematics; it is a specific signature of smooth, differentiable functions. If the underlying physics were to involve sharp, jagged features, the statistical behavior of the peaks would look completely different. This distinction helps physicists understand that the chaotic behavior seen in string scattering is rooted in the smooth, continuous nature of the theory, rather than in some hidden randomness.

By connecting the dots between the local geometry of the amplitude and the global statistics of the peaks, the researchers have identified a new class of universal behavior. They showed that the repulsion is not a result of the peaks being eigenvalues of a random matrix, as was previously hypothesized in some contexts. Instead, it is a local event driven by the derivatives of the function. The team's work bridges the gap between abstract mathematical theory and concrete numerical evidence, showing that the "chaos" in string scattering is a predictable, geometric inevitability. Their findings suggest that if we look at other physical theories that produce smooth scattering amplitudes, we should expect to see this same quartic repulsion. It is a reminder that even in the most complex and chaotic-looking systems, there are simple, underlying geometric rules that dictate how the pieces fit together. The study does not claim to solve the entire mystery of quantum chaos, but it provides a clear, local mechanism for one of its most striking features, turning a statistical observation into a geometric certainty.

Technical Summary: The Origin of Amplitude Chaos and Strings

Problem Statement
The paper investigates the statistical mechanics of scattering amplitudes in highly excited string theory, specifically addressing the phenomenon of "peak repulsion." In closed Hamiltonian systems, quantum chaos is often diagnosed by level repulsion between eigenvalues, typically described by Random Matrix Theory (RMT). However, scattering amplitudes do not possess a discrete energy spectrum; instead, they exhibit a complex angular dependence with local extrema (peaks). Previous studies on highly excited string amplitudes (specifically the three-point amplitude with one highly excited string and two tachyons, known as HTT) observed that neighboring peaks avoid being arbitrarily close, suggesting a chaotic structure.

A critical gap exists in the theoretical understanding of this phenomenon: while RMT explains eigenvalue repulsion via the Jacobian of the transformation from matrix entries to eigenvalues, string amplitude peaks are local extrema of exact, deterministic functions derived from integer partitions of the excitation level $N$ . There is no underlying random matrix being diagonalized, and the ensemble is generated by sampling known partitions, not random matrix entries. The authors ask: What local mechanism forces two physical peaks to avoid contact, and does this repulsion depend on the specific details of string theory or represent a broader universality class?

Methodology
The authors employ a dual approach combining rigorous analytical derivation with extensive numerical verification.

Analytical Framework:
- Direct Peak Collision Analysis (Section 2): The authors model the amplitude $A(\alpha)$ as a smooth, real-valued function. They analyze the conditions required for two adjacent peaks to approach a separation $S \to 0$ . By expanding the amplitude in a Taylor series around the midpoint between two peaks, they derive constraints on the coefficients (derivatives) $A(0), A'(0), A''(0), A'''(0)$ . They assume the ensemble of amplitudes samples these derivatives from a smooth, non-degenerate probability density.
- Equivalent Three-Roots Description (Section 3): Recognizing that peaks lie between consecutive roots of the amplitude, the authors reformulate the problem in terms of three consecutive roots $z_0, z_1, z_2$ . They derive the joint probability density of these roots, identifying a Vandermonde-like determinant factor $|\Delta_3(z)| = |(z_1-z_0)(z_2-z_1)(z_2-z_0)|$ arising from the change of variables from polynomial coefficients to roots. This factor is shown to suppress the probability of roots (and thus the peaks between them) approaching each other.
Numerical Experiments (Section 4):
- The authors compute the exact HTT amplitudes for various excitation levels $N$ (ranging from 320 to 1280) by sampling integer partitions uniformly.
- They implement a numerically stable algorithm to locate exact peaks within zero-free cells defined by the amplitude's roots.
- They measure the nearest-neighbor peak spacing distribution $p_{nn}(S)$ and extract the contact exponent $\beta$ defined by $p_{nn}(S) \sim S^\beta$ as $S \to 0$ .
- They test the robustness of the results by varying the fitting window and comparing direct peak spacings ( $s$ ) with spacings derived from the three-root approximation ( $r$ ).
Generalization and Sensitivity Tests (Section 5):
- Toy Models: The authors test the theory on a "toy" amplitude constructed from integer partitions and sine functions, and on random Fourier series with coefficients drawn from various distributions (Gaussian, Uniform, Laplace, Student-t, Smooth Compact).
- Differentiability Tests: They introduce a tunable cusp model $A_\gamma(x) \sim |x|^\gamma$ to explicitly vary the local differentiability of the amplitude function and observe the effect on the contact exponent.

Key Contributions and Results

Derivation of Quartic Repulsion: The primary analytical result is the derivation of a quartic contact law for the peak spacing distribution:
$p_{nn}(S) \sim S^4 \quad \text{as} \quad S \to 0$
This result is obtained through two independent methods:
1. Direct integration over the joint probability density of the first four Taylor coefficients, showing that the volume of the phase space allowing two peaks to be within distance $S$ scales as $S^4$ .
2. Analysis of the three-root configuration, where the Vandermonde factor from the root-to-coefficient transformation yields a probability density scaling as $S^4$ (specifically $uv(u+v) \sim S^4$ for small gaps).
Numerical Confirmation in String Theory: Numerical simulations of the HTT amplitude confirm the analytical prediction. For large $N$ , the measured contact exponent $\beta$ consistently converges to 4 (within statistical errors and window variations), validating the quartic repulsion in the context of string theory. The ratio between exact peak spacing and the root-derived spacing is found to be $\approx 0.99$ , confirming the validity of the three-root approximation.
Universality Beyond Strings: The study demonstrates that the quartic repulsion is not unique to string theory.
- Toy models with integer partitions and sine functions exhibit the same $\beta \approx 4$ .
- Random Fourier series with coefficients drawn from diverse smooth distributions (including heavy-tailed and compactly supported laws) all yield $\beta \approx 4$ .
- This establishes a universality class for amplitude ensembles: any ensemble of sufficiently smooth ( $C^3$ ), non-degenerate analytic functions drawn from a smooth probability distribution will exhibit quartic peak repulsion.
Role of Differentiability: The paper identifies the boundary of this universality. By reducing the local differentiability of the amplitude function (using the cusp model with parameter $\gamma$ ), the contact exponent changes continuously according to the relation:
$\beta(\gamma) = 2\gamma - 2$
For $\gamma = 3$ (analytic cubic), $\beta = 4$ . For lower $\gamma$ , the repulsion becomes weaker (sub-quartic). This confirms that the quartic law is a direct consequence of the analytic structure (specifically the existence of continuous third derivatives) and the non-degeneracy of the derivative sampling.

Significance and Claims
The paper claims to resolve the mechanism behind peak repulsion in string scattering amplitudes without invoking Random Matrix Theory as the fundamental cause. Instead, it attributes the repulsion to the local geometric constraints imposed by the smoothness of the amplitude function and the statistical sampling of its derivatives.

Mechanism: Peak repulsion arises because the condition for two peaks to be arbitrarily close requires a highly constrained, low-probability configuration of the amplitude's local derivatives (specifically, the first three derivatives must vanish or scale in a specific way).
Universality: The quartic exponent is a generic feature of smooth amplitude ensembles, not a specific signature of string theory. This suggests that similar repulsion statistics should be observable in other theories with smooth scattering amplitudes.
Limitations: The authors are modest regarding the global structure of the spacing distribution. While the small-spacing limit is fixed by local smoothness (yielding $S^4$ ), the behavior at moderate and large spacings depends on the specific ensemble details and is not necessarily described by a single Gaussian RMT ensemble (e.g., GOE). The paper clarifies that the absence of a global RMT fit in string amplitudes does not contradict the local quartic repulsion.

In summary, the work defines a local universality class for amplitude chaos, identifying smoothness and non-degeneracy as the necessary and sufficient conditions for quartic peak repulsion, and provides a clear analytical and numerical framework for distinguishing this local behavior from global spectral statistics.

More like this