Multicriticality and Scaling: Mellin Spectral Theory, and the Decoupling of Geometric and Spectral Exponents

This paper develops a spectral theory for scale-invariant operators on the multiplicative half-line using Mellin transforms to demonstrate that geometric and spectral exponents are fundamentally decoupled, providing a precise mathematical characterization of multicriticality where their inequality signals multiple independent scaling dimensions.

The Gist

The Big Picture: Two Different "Rulers" for the Same Object

Imagine you are looking at a complex pattern, like a snowflake or a network of connections between people. In physics and math, when a system is "critical" (meaning it's on the edge of a major change, like water turning to ice), it often looks the same no matter how much you zoom in or out. This is called scale invariance.

Usually, scientists assume there is just one rule that describes how this pattern shrinks or grows. This paper argues that there are actually two different rulers measuring the same thing, and they often give different answers.

1. The Geometric Ruler (The "Envelope"): This measures the overall shape or the "skin" of the pattern. It tells you how the whole thing scales up or down.
2. The Spectral Ruler (The "Inner Rhythm"): This measures the internal vibrations or the specific "notes" the pattern plays. It tells you how the strength of those internal parts decays.

The paper's main discovery is that these two rulers are decoupled. They don't have to agree. When they disagree, the system is "multicritical" (having multiple complex scaling behaviors). When they agree, it's a simple critical point.

---

The Mathematical Machine: The "Mellin" Lens

To prove this, the authors built a special mathematical machine called the Mellin Transform.

The Analogy: The Prism
Think of a beam of white light hitting a prism. The prism splits the light into a rainbow of colors.

• In this paper, the "white light" is a complex mathematical function (a kernel) that describes how different points in a system interact.
• The "prism" is the Mellin Transform.
• When you shine the function through the prism, it doesn't just split into colors; it splits into pure tones (eigenfunctions).

The paper shows that for any system that looks the same at different scales, this prism reveals a very specific structure:

• The Shape: The function is made of a "power-law envelope" (a smooth, predictable curve that gets smaller as you go out) multiplied by a "shape function" (the specific details of the pattern).
• The Result: The prism separates these two. The envelope is determined by the Geometric Exponent ($a$), and the details are determined by the Spectral Exponent ($b$).

The "Lorentzian" Surprise

The authors tested this with a specific, simple pattern (a kernel involving an exponential decay).

• What they expected: They thought the internal "notes" (eigenvalues) would follow a simple power-law rule, just like the outer shape.
• What they found: The internal notes followed a Lorentzian shape (a specific bell-curve-like shape often seen in physics, like the resonance of a tuning fork).
• The Consequence: Because the internal notes follow a Lorentzian curve, the "Spectral Exponent" ($b$) calculated from them is different from the "Geometric Exponent" ($a$) of the outer shape.

The Takeaway: Just because a system looks like it scales in a certain way on the outside doesn't mean its internal parts scale the same way. They are independent.

The Lattice Trap: Why You Can't Count on Discrete Steps

The paper also addresses a common problem: What happens if you try to do this math on a grid of integers (like a computer screen made of pixels) instead of a smooth, continuous line?

The Analogy: The Broken Mirror
Imagine trying to take a perfect, smooth reflection of a mountain in a mirror made of jagged, discrete tiles.

• The authors proved a "Collapse Theorem." If you try to force the rules of scale invariance onto a discrete grid of integers, the math breaks down.
• Instead of having many different "modes" or "vibrations," the grid forces all the eigenvectors (the patterns) to collapse into a single, identical shape. It's like trying to play a symphony on a piano where every key produces the exact same note.
• The Solution: You must move to the "continuum" (smooth numbers) to see the full, rich spectrum of behavior. The discrete grid is just a rough, low-resolution sampling of the smooth reality.

Why This Matters for "Multicriticality"

In the language of the paper:

• Simple Criticality: The Geometric Exponent ($a$) equals the Spectral Exponent ($b$). The system is simple; the outside and inside scale together.
• Multicriticality: The Geometric Exponent ($a$) does not equal the Spectral Exponent ($b$). The system is complex; it has multiple independent scaling dimensions.

The paper provides a precise mathematical definition for this complexity: Multicriticality is simply the condition where $a \neq b$.

Summary of the "Real World" Claims

The paper claims that:

1. Scale-invariant systems can be mathematically split into a "geometric envelope" and a "spectral shape."
2. These two parts are independent; the shape of the envelope does not dictate the decay of the internal spectrum.
3. Trying to analyze this on a discrete grid (like a computer matrix) causes a mathematical "collapse" where all patterns look the same, which is why we need continuous math to understand the true behavior.
4. The difference between the geometric scaling and the spectral scaling is the rigorous definition of a "multicritical" system.

The paper does not claim to diagnose specific diseases, predict stock market crashes, or solve biological problems directly. It strictly provides the mathematical foundation (the "rulers" and the "prism") that could be used to understand such systems, noting that the ratio of these two exponents ($a/b$) measures the degree of complexity.

Technical Summary

Technical Summary: Multicriticality and Scaling via Mellin Spectral Theory

Problem Statement
The paper addresses the mathematical characterization of scale-invariant systems, specifically focusing on the structure of symmetric operators whose kernels satisfy a scaling relation $M(kx, ky) = k^{-a}M(x, y)$. While scale invariance is a hallmark of critical systems in the Renormalization Group (RG) framework, existing literature often conflates the geometric scaling of the system's envelope with the spectral decay of its eigenvalues. The authors seek to rigorously define the spectral theory of such operators on the multiplicative half-line $(\mathbb{R}^+, dx/x)$ and investigate whether the geometric exponent governing the kernel's envelope is necessarily identical to the effective spectral exponent observed in finite-dimensional truncations. A secondary problem is the degeneracy that arises when attempting to formulate self-similarity directly on a discrete integer lattice.

Methodology
The authors develop a spectral theory based on the Mellin transform, which serves as the Fourier transform on the multiplicative group $(\mathbb{R}^+, \times)$.

1. Hilbert Space Formulation: The operators are defined on the Hilbert space $H = L^2(\mathbb{R}^+, dx/x)$, utilizing the Haar measure of the multiplicative group.
2. Kernel Factorization: The authors prove that any symmetric scale-invariant kernel of order $a$ must factorize into a universal power-law envelope $(xy)^{-a/2}$ and a shape function $F(x/y)$ dependent only on the ratio of arguments.
3. Diagonalization: Using the Mellin transform, the integral operator is diagonalized. The generalized eigenfunctions are identified as Mellin modes $\psi_\omega(x) = x^{-a/2 + i\omega}$, and the eigenvalues correspond to the Mellin multiplier $\tilde{F}(\omega)$.
4. Discrete Analysis: The paper analyzes the implications of imposing discrete self-similarity on the integer lattice $\mathbb{N}$, proving a theorem regarding eigenvector collapse.
5. Numerical Verification: The theory is tested using a specific kernel $M(x, y) = c(xy)^{-a/2} \rho^{|\ln(x/y)|}$, where the shape function is an exponential decay in log-space. Numerical simulations on finite $N \times N$ truncations are used to compare the geometric exponent $a$ with the effective spectral exponent $b$ derived from eigenvalue decay.

Key Contributions and Results

• Kernel Factorization Theorem: The paper establishes that scale-invariant kernels are completely characterized by the form $M(x, y) = (xy)^{-a/2} F(x/y)$. This separates the geometric scaling (the envelope) from the correlation structure (the shape function).
• Mellin Diagonalization: The authors demonstrate that the Mellin transform diagonalizes these operators, yielding a continuous spectrum parametrized by $\omega \in \mathbb{R}$. The eigenvalues are given by the Mellin multiplier $\tilde{F}(\omega)$.
• Decoupling of Exponents: A central result is the demonstration that the geometric exponent $a$ (from the kernel envelope) and the spectral exponent $b$ (the effective power-law index of the discrete eigenvalue spectrum, $\lambda_n \sim n^{-b}$) are generically distinct.
  • In the specific example where $F(t) = c \rho^{|\ln t|}$, the Mellin multiplier is a Lorentzian function, not a power law.
  • Consequently, the discrete eigenvalues of a finite truncation exhibit an approximate power-law decay with an exponent $b$ that depends on the width of the Lorentzian ($\sigma = -\ln \rho$) and the matrix size $N$, rather than the geometric exponent $a$.
• Eigenvector Collapse on the Lattice: The paper proves that imposing a strict discrete self-similarity condition on the integer lattice forces all eigenvectors to be proportional to a single sequence ($v_m^{(n)} \propto m^{-a/2}$), resulting in a rank-one matrix. This necessitates the continuum formulation, where eigenvectors are generalized distributions (analogous to plane waves) and the spectrum is continuous.
• Characterization of Multicriticality: The authors define multicriticality as the condition where $a \neq b$.
  • If $a = b$, the system corresponds to a simple critical fixed point in the RG framework.
  • If $a \neq b$, the system possesses multiple independent scaling dimensions, signaling multicriticality. The ratio $a/b$ serves as a quantitative measure of this decoupling.
• Diagnostic Implications: The paper clarifies that eigenvector self-similarity is a generic property of smooth covariance operators and is not a unique signature of criticality. Instead, the robust diagnostic for critical organization is the power-law scaling of the eigenvalues (spectral exponent), not the rescaling of eigenvectors.

Significance and Claims
The paper claims to provide a rigorous mathematical foundation for the concept of multicriticality, distinguishing it from simple criticality through the decoupling of geometric and spectral scaling dimensions. By utilizing Mellin spectral theory, the authors resolve the ambiguity often found in finite-dimensional matrix analyses where the effective spectral exponent $b$ is mistakenly assumed to be identical to the geometric exponent $a$.

The work establishes that:

1. The continuum formulation is necessary to avoid the spectral degeneracy (rank-one collapse) inherent in discrete lattice formulations of self-similarity.
2. The equality $a = b$ is a specific condition for a simple RG fixed point, while $a \neq b$ is the structural signature of systems with multiple scaling dimensions.
3. The Lorentzian form of the Mellin multiplier (arising from exponential decay in log-space) naturally leads to a decoupling of exponents in finite truncations, providing a concrete mechanism for observing multicriticality in numerical data.

The authors note that while the framework is developed for scale-invariant operators, it offers a precise tool for analyzing empirical correlation matrices in complex systems (financial, biological, neural) to detect proximity to criticality and quantify the degree of multicriticality. They also draw a structural parallel between the mode-dependent scaling factors in their generalized theory and the frequency distributions in the Kuramoto model of collective synchronization, suggesting that the Lorentzian's role in both contexts (enabling exact dimensional reduction) is a deep mathematical connection rather than a coincidence.