🔢 mathematics

Which Saddles Contribute? The South-East Rule for Multidimensional Integrals

This paper introduces a simple geometric "South-East" rule algorithm that determines which critical points contribute to the asymptotic evaluation of multidimensional integrals with exponential integrands, thereby eliminating the need for complex flow computations required by traditional Picard-Lefschetz methods and offering a systematic approach to identifying instanton contributions in real-time path integrals.

Original authors: Inês Aniceto, Job Feldbrugge, Christopher J. Howls

Published 2026-06-29

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Inês Aniceto, Job Feldbrugge, Christopher J. Howls

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

The Big Picture: Navigating a Foggy Mountain Range

Imagine you are trying to calculate the total "energy" of a complex system (like a wave of light or a quantum particle) by adding up contributions from every possible path it could take. In math, this is done using a multidimensional integral.

The problem is that the landscape of possibilities is a massive, foggy mountain range with millions of peaks and valleys. Most of these paths cancel each other out because they are moving in opposite directions (destructive interference). Only a few specific "peaks" (called critical points or saddles) actually contribute to the final answer.

For decades, mathematicians and physicists have struggled with a simple question: Which of these millions of peaks actually matter?

Traditionally, to find the right peaks, you had to simulate a river flowing down the mountain from every single point to see where it ended up. This is like trying to map a continent by walking every single inch of it. It's slow, difficult, and gets impossible as the mountain gets bigger (more dimensions).

The New Solution: The "South-East Rule"

This paper introduces a simple, geometric shortcut called the South-East Rule. Instead of walking the whole mountain, you can look at a specific map (called the Borel plane) and instantly know which peaks matter.

Here is how the analogy works:

1. The Peaks and the Map

Imagine all the important peaks in your mountain range are plotted on a 2D map.

Some peaks are Real (they exist in our normal world).
Some peaks are Complex (they exist in a hidden, mathematical dimension).

The authors discovered that these peaks are connected by invisible "roads" called adjacency lines. If two peaks are connected, they can influence each other. If they aren't connected, they are isolated islands.

2. The "South-East" Direction

The paper's main discovery is a simple rule for determining which complex peaks become "active" contributors:

The Setup: Imagine the "Real" peaks are your starting point.
The Rule: Look at the map. If a complex peak is located to the South-East of a relevant Real peak (meaning it is lower and to the right), and there is a direct road (adjacency) connecting them, then that complex peak contributes to the final answer.
The Logic: Think of it like water flowing. If you are at a high point (Real peak), water can only flow "downhill" and "eastward" to reach a lower point. If a complex peak is in that specific South-East direction and connected by a road, the "influence" flows to it, making it relevant. If it's North-West, or if the road is blocked, it stays irrelevant.

3. Why This is a Big Deal

No More Walking: You don't need to simulate the complex "flow" of water down the mountain (which is what the old method required). You just look at the coordinates of the peaks and check if one is South-East of the other.
Works in Any Dimension: Whether you are looking at a 1D line, a 2D surface, or a 100D hyper-space, this geometric rule holds true.
Handles the "Unbounded" Problem: Sometimes the mountain doesn't have a bottom (the math goes to negative infinity). The authors show that by adding a temporary "regulator" (like putting a floor under the mountain), you can apply the rule, and then remove the floor to get the correct answer.

Real-World Examples Used in the Paper

The authors tested this rule on several classic mathematical shapes (called "catastrophes") to prove it works:

The Cusp (Pearcey Integral): A shape that looks like a sharp point. They showed how the rule correctly identifies which peaks matter as you change the shape's parameters.
The Swallowtail: A more complex shape with a tail. They demonstrated that even when the mountain is "unbounded" (no bottom), the rule works if you use their regulator trick.
The Hyperbolic Integral: A 2D surface. They used the rule to find complex peaks that traditional methods might miss, and their predictions matched computer simulations perfectly.

Why Physicists Care (According to the Paper)

The paper connects this math to Quantum Mechanics and Path Integrals.

In quantum physics, particles take "all possible paths" to get from A to B.
Most paths cancel out. Only specific "instantons" (special paths) matter.
Finding these instantons in "real-time" (not just in imaginary time) has been a huge headache for physicists, often leading to mathematical infinities (the "sign problem").
The authors suggest that the South-East Rule provides a systematic way to find these relevant instantons without getting stuck in the math. It offers a potential way to fix the "Wick rotation" problem (a method used to make quantum math easier) when the energy of the system isn't bounded from below.

Summary

The paper says: "Stop trying to walk the whole mountain to find the right path. Just look at the map. If a hidden peak is connected to a real peak and sits in the South-East direction, it counts. If not, ignore it."

This simple geometric check replaces a massive, complex calculation, making it much easier to solve difficult problems in physics and mathematics.

Technical Summary: The South-East Rule for Multidimensional Integrals

Problem Statement
The paper addresses a fundamental challenge in the asymptotic evaluation of multidimensional oscillatory integrals of the form $\Psi = (\frac{k}{2\pi i})^{d/2} \int_{\mathbb{R}^d} e^{ikf(x)} dx$ , where $f$ is an analytic function and $k$ is a large parameter. While the asymptotic behavior of such integrals is governed by critical points (saddles) of the phase function $f$ , a persistent difficulty lies in determining which critical points—both real and complex—actually contribute to the integral's asymptotic expansion.

In one dimension ( $d=1$ ), this is resolved by geometrically tracing steepest descent contours. However, for higher dimensions ( $d > 1$ ), the standard approach relies on Picard-Lefschetz theory, which requires deforming the real integration domain $\mathbb{R}^d$ into the complex space $\mathbb{C}^d$ along steepest descent manifolds. Identifying the relevant critical points in this framework necessitates solving the downward flow partial differential equation ( $\dot{x} = -\nabla \text{Re}[if(x)]$ ) to find manifolds connecting the original domain to critical points. This procedure becomes computationally intractable as the dimension $d$ increases. Furthermore, existing methods often assume the relevant critical points are known a priori, lacking a general, practical criterion for their identification.

Methodology
The authors propose a dimension-independent algorithm that bypasses the need to compute flow manifolds in $\mathbb{C}^d$ . Instead, the method synthesizes three key elements:

Borel Plane Geometry: The integral is transformed into a single-complex-dimensional Borel plane representation using the variable $t = -if(x)$. In this plane, critical points map to branch point singularities, and the original integration domain maps to a specific contour (typically along the imaginary axis).
Resurgence and Adjacency: The authors utilize the theory of resurgence, specifically the "adjacency" relations between critical points. Two critical points are "adjacent" if their asymptotic series are linked via Stokes phenomena, which geometrically corresponds to the ability of their associated branch points to "see" each other on the Riemann sheets of the Borel plane Jacobian. This adjacency can be determined algebraically via hyperasymptotics or numerically via Borel-Padé approximants.
The "South-East" Rule: The core innovation is a geometric rule applied to the adjacency graph in the Borel plane. By rotating the phase of the integral parameter $k$ (effectively rotating the integration domain and critical points in the Borel plane), the authors track Stokes transitions. They demonstrate that a complex critical point becomes relevant (i.e., contributes to the integral) if, during the rotation from a non-oscillatory regime ( $\theta = \pi/2$ ) back to the original oscillatory regime ( $\theta = 0$ ), it is "activated" by a relevant critical point via an adjacency edge oriented in the "South-East" direction (slope $\theta_{jk} \in [-\pi/2, 0]$ ).

The algorithm distinguishes between two cases:

Bounded Exponentials: Where $f(x)$ is bounded from below. The algorithm starts with real critical points as relevant at $\theta = \pi/2$ and tracks Stokes transitions as $\theta \to 0$ .
Unbounded Exponentials: Where $f(x)$ is unbounded (e.g., Airy or Swallowtail integrals). The authors introduce a smooth analytic regulator to bound the potential, apply the algorithm, and then take the limit as the regulator vanishes. This allows the identification of relevant complex critical points that would otherwise be missed or cause divergence in naive Wick rotations.

Key Contributions and Results

Algorithmic Identification: The paper provides "Algorithm 1," a step-by-step procedure to determine intersection numbers ( $n_j$ ) and thus the relevance of critical points for arbitrary dimensions $d \ge 1$ . This eliminates the need for solving high-dimensional flow equations.
Validation via Examples: The rule is tested on canonical diffraction integrals, including the Pearcey (cusp), Swallowtail, and Hyperbolic integrals. In these cases, the algorithm correctly identifies the set of relevant real and complex saddles across different regions of the parameter space (separated by caustics and Stokes lines).
Numerical Verification: For the Hyperbolic integral, the authors compare the asymptotic approximation derived using the South-East rule against numerical integration. The inclusion of complex critical points identified by the rule significantly improves the accuracy of the approximation, particularly for smaller values of $k$ , confirming the rule's predictive power.
Handling Unbounded Potentials: The paper demonstrates how to handle integrals with unbounded exponents (where standard Wick rotation fails) by using regulators and tracking the evolution of relevance through the Borel plane, effectively solving the "sign problem" for specific analytic continuations.

Significance and Claims
The authors claim this work represents a "significant step towards a systematic way of identifying instanton contributions in real-time path integrals." By linking the algebraic structure of resurgence (adjacency) directly to the geometric relevance of critical points, the paper offers a practical tool for:

Resolving Wick Rotation Issues: Providing a rigorous framework for defining path integrals where the potential is not bounded from below, a problem central to the conformal factor problem in quantum gravity and the sign problem in many-particle systems.
High-Dimensional Integration: Offering a computationally efficient alternative to numerical evaluation for high-dimensional oscillatory integrals, which are notoriously expensive to compute.
Quantum Physics Applications: Enabling the identification of relevant complex classical paths (instantons) in real-time quantum mechanics and quantum field theory, which are essential for non-perturbative effects but difficult to isolate using traditional methods.

The paper concludes that the South-East rule, combined with hyperasymptotic or Borel-Padé resummation, allows for the recovery of highly oscillatory integrals without direct numerical integration, marking a potential shift in how asymptotic expansions are derived for multidimensional systems.