Integral Means Spectrum for the Random Riemann Zeta Function

This paper demonstrates that the integral means spectrum of the primitive of the randomized Riemann zeta-function and holomorphic Gaussian multiplicative chaos almost surely follows the Kraetzer conjecture, despite neither function being injective.

The Gist

Imagine you are standing on the edge of a vast, chaotic ocean. This ocean is the Riemann Zeta Function, a famous mathematical object that holds the secrets to how prime numbers are distributed. It's wild, unpredictable, and full of towering waves and deep troughs.

For decades, mathematicians have tried to measure the "roughness" of this ocean. They want to know: if you zoom in on a tiny patch of the water, how wild does it get? Does it look like a calm lake, or like a stormy sea with jagged peaks?

This paper, written by Duplantier, Gayrard, and Saksman, is like a new set of high-tech sensors dropped into this ocean. They discover that the "roughness" of the Zeta function follows a very specific, beautiful pattern that was predicted 30 years ago but never proven until now.

Here is the story of their discovery, broken down into simple concepts.

1. The Random Ocean (The Random Zeta Function)

The actual Zeta function is too complicated to study directly in its wild state. So, the authors create a simulation. They build a "Random Zeta Function."

Think of this like a video game version of the ocean. Instead of the real, complex physics, they use a recipe involving random numbers (specifically, random choices for every prime number).

• The Recipe: They take every prime number (2, 3, 5, 7, 11...) and assign it a random direction (like a coin flip).
• The Result: When they combine these, they get a function that behaves statistically just like the real Zeta function near its most critical zone (the "critical strip"). It's a perfect stand-in to test theories without getting lost in the real function's complexity.

2. Measuring the Storm (The Integral Means Spectrum)

The authors want to measure the "intensity" of this random ocean. They define a tool called the Integral Means Spectrum.

Imagine you are a fisherman casting a net around a specific area of the ocean.

• You don't just measure the height of one wave.
• You measure the average power of all the waves in that net, but you do it for different "magnifications."
• If you look at the average height, it's one thing. If you look at the average of the squares of the heights (emphasizing the huge, dangerous waves), it's another. If you look at the cubes, you are only seeing the most extreme, terrifying spikes.

The "Spectrum" is a graph that tells you: "As we zoom in closer and closer to the edge of the ocean, how does the average power of these waves grow?"

3. The Prediction: The "Kraetzer Curve"

Thirty years ago, a mathematician named Kraetzer made a bold guess about what this graph should look like for any random, complex shape (not just the Zeta function). He called it the Universal Spectrum.

His guess was a simple, elegant curve:

• For small magnifications: The roughness grows like a parabola (a smooth U-shape). It's gentle.
• For huge magnifications: The roughness grows like a straight line. It hits a "speed limit" where the waves can't get any wilder relative to the zoom.

This curve is famous in the world of "fractals" (shapes that look the same at every zoom level). It's the "Gold Standard" for randomness.

4. The Discovery: It's a Match!

The authors did the hard math (using tools from probability theory and number theory) to calculate the spectrum for their Random Zeta Function.

The Result: The graph they calculated fits Kraetzer's prediction perfectly.

• The "small magnification" part is exactly the parabola ($\beta^2/4$).
• The "huge magnification" part is exactly the straight line ($|\beta| - 1$).

This is huge because it proves that the Riemann Zeta function, despite being a number-theoretic object, behaves exactly like a "universal" random storm. It connects the world of prime numbers to the world of random physics.

5. The Twist: It's Not a Map (Non-Injectivity)

There is a catch. In math, when we study these shapes, we often hope they are "one-to-one" (injective). This means that if you draw a line on the map, it doesn't cross over itself. It's a clean, non-overlapping shape.

The authors checked if their Random Zeta function was a "clean" map.

• The Verdict: No.
• The Analogy: Imagine you are drawing a path on a piece of paper. You expect to draw a line that never touches itself. But when you draw this Random Zeta path, it gets so tangled and chaotic that it crosses over itself like a ball of yarn.
• Why it matters: Even though the statistics of the storm (the spectrum) are perfect and follow the universal law, the actual shape of the function is messy and tangled. It's a beautiful storm, but it's a messy one.

6. The Connection to "Chaos" and Gravity

The paper also links this to Gaussian Multiplicative Chaos (GMC).

• The Metaphor: Imagine a fog that isn't just random, but where the density of the fog depends on the fog itself. It's a self-amplifying chaos.
• The authors show that the Random Zeta function is essentially a version of this "Chaos Fog."
• This connects the Zeta function to Liouville Quantum Gravity, a theory physicists use to describe how space-time might look if it were made of tiny, fluctuating quantum bubbles.

Summary

In simple terms, this paper says:

1. We built a random version of the famous Riemann Zeta function.
2. We measured how "wild" it gets as we zoom in.
3. It turns out to follow a 30-year-old prediction perfectly, proving that the Zeta function shares the same fundamental "roughness" as other random systems in nature.
4. However, the shape itself is tangled and messy, not a clean, simple curve.

It's a victory for the idea that randomness has a hidden order, even in the most mysterious corners of mathematics.

Technical Summary

1. Problem Statement and Context

The paper investigates the integral means spectrum of the primitive (antiderivative) of the randomized Riemann zeta function, denoted as $\zeta_{\text{rand}}(s)$. This function, introduced by Bagchi, serves as a probabilistic model for the statistical behavior of the actual Riemann zeta function $\zeta(s)$ under random vertical shifts in the critical strip ($1/2 < \sigma \le 1$).

The central object of study is the function $F(s) = \int_1^s \zeta_{\text{rand}}(z) dz$. The authors aim to determine the asymptotic growth of the integral moments of its derivative, $F'(s) = \zeta_{\text{rand}}(s)$, as the real part $\sigma$ approaches the critical line $1/2$ from the right. Specifically, they seek to calculate the spectrum function $f(\beta)$ defined by:
$$\lim_{\sigma \to 1/2^+} \frac{\log \int_0^1 |\zeta_{\text{rand}}(\sigma + ih)|^\beta dh}{\log((\sigma - 1/2)^{-1})} = f(\beta)$$
This problem is deeply connected to:

• Universal Spectra in Complex Analysis: The Kraetzer Conjecture, which proposes a specific form for the universal integral means spectrum of univalent (injective) functions in the unit disk.
• Statistical Physics: The Random Energy Model (REM) and Gaussian Multiplicative Chaos (GMC), where similar scaling exponents and "freezing transitions" occur.
• Injectivity: A crucial question is whether the primitive $F$ is injective (univalent), as the Kraetzer conjecture specifically applies to univalent functions.

2. Methodology

The authors employ a dual approach, combining probabilistic number theory with stochastic analysis of Gaussian Multiplicative Chaos.

A. Probabilistic Number Theory Approach (Theorem 1.1)

To prove the result for the random zeta function in the half-plane, the authors utilize:

1. Multiscale Decomposition: They decompose the logarithm of the random zeta function into a sum over dyadic scales of prime numbers. The process $X_h^\sigma$ is rewritten as a sum of independent increments $Y_h^\sigma(k)$ corresponding to primes in intervals $(2^{k-1}, 2^k]$.
2. Prime Number Theorem (PNT): The PNT is used to estimate the variance and covariance of these increments, controlling short and long-range correlations.
3. Kistler's Multiscale Second Moment Method: Adapted from the study of log-correlated fields, this method is used to handle the correlations between the process at different points $h$ and $h'$. It allows for precise estimation of the "free energy" (the limit of the log-moments).
4. Berry-Esseen Theorems: Used to establish the Gaussian nature of the partial sums of the increments, ensuring the validity of large deviation estimates.
5. Paley-Zygmund Inequality: Applied to prove lower bounds on the measure of "high points" (where the process exceeds a certain threshold), which is essential for establishing the lower bound of the spectrum.

B. Gaussian Multiplicative Chaos (GMC) Approach (Theorem 1.2 & Alternative Proof)

The authors also study a unit disk analogue where the derivative is defined via a holomorphic GMC:
$$F'(z) = \exp\left( \sum_{n=1}^\infty \frac{G_n z^n}{\sqrt{n}} \right)$$
where $G_n$ are independent complex Gaussians.

1. GMC Theory: They leverage established results on the convergence of GMC measures and the properties of log-correlated fields.
2. Convexity Arguments: Using the log-convexity of the $L^\beta$ norms (Hadamard's three-lines theorem analogue), they extend results from the subcritical regime ($\beta < 2$) to the supercritical regime ($\beta \ge 2$).
3. Coupling: They utilize a coupling result from prior work (Arguin, Dubédat, et al.) to show that the random zeta function converges to a holomorphic GMC distribution, providing an alternative derivation of the spectrum.

C. Injectivity Analysis (Propositions 1.3 & 1.4)

To address the univalence question, the authors apply the Becker-Pommerenke injectivity criterion. They analyze the quantity $(1-|z|^2)|F''(z)/F'(z)|$ (or its half-plane equivalent) and show that it diverges almost surely as one approaches the critical line, proving that the primitive is not injective.

3. Key Results

The Main Theorem: The Kraetzer Spectrum

The primary result is that the integral means spectrum of the primitive of the random zeta function (and its GMC analogue) is almost surely given by the extended Kraetzer conjecture:

$$f(\beta) = \begin{cases} \frac{1}{4}|\beta|^2 & \text{if } |\beta| \le 2 \\ |\beta| - 1 & \text{if } |\beta| \ge 2 \end{cases}$$

• Phase Transition: The spectrum exhibits a "freezing transition" at $|\beta| = 2$.
  • For $|\beta| \le 2$, the growth is quadratic ($\beta^2/4$), dominated by the typical fluctuations of the field (the "high-temperature" phase).
  • For $|\beta| \ge 2$, the growth is linear ($|\beta|-1$), dominated by the extreme values of the field (the "low-temperature" spin-glass phase).
• Universality: This result confirms that the random zeta function belongs to the same universality class as the Random Energy Model and holomorphic GMC, sharing the exact same multifractal spectrum.

Non-Injectivity

Despite the spectrum matching the form conjectured for univalent functions, the authors prove that the primitive $F$ is almost surely not injective in any neighborhood of the critical line (or in the unit disk).

• This is a significant finding because it demonstrates that the Kraetzer spectrum form is not exclusive to univalent functions; it arises naturally in non-injective random analytic functions driven by log-correlated noise.

Convergence

The convergence of the spectrum is established both almost surely and in $L^q$ mean for any $q \ge 0$.

4. Significance and Implications

1. Resolution of a Conjecture in a Random Setting: The paper provides the first rigorous proof that a specific random analytic function (the primitive of $\zeta_{\text{rand}}$) almost surely realizes the Kraetzer spectrum. This validates the conjecture in a probabilistic context, even though the function is not univalent.
2. Bridge Between Number Theory and Statistical Physics: The work rigorously connects the statistical properties of the Riemann zeta function (specifically its moments and maxima) to the theory of Gaussian Multiplicative Chaos and the Random Energy Model. It confirms that the "freezing transition" observed in the zeta function's moments is a manifestation of the same phase transition seen in spin glasses.
3. Clarification of Univalence and Spectra: By showing that the Kraetzer spectrum appears for non-injective functions, the authors clarify that the form of the spectrum is a property of the underlying log-correlated field's extreme value statistics, rather than a strict consequence of geometric univalence.
4. Methodological Advancement: The paper successfully adapts multiscale second-moment methods (originally developed for branching random walks and BBM) to the context of the Riemann zeta function, offering a robust toolkit for analyzing the fine structure of random zeta functions.

In summary, the paper establishes that the integral means spectrum of the random Riemann zeta function follows the universal Kraetzer law, characterized by a quadratic-to-linear transition at $|\beta|=2$, and demonstrates that this universal behavior is intrinsic to the log-correlated nature of the field, independent of the injectivity of the function.