Effective equidistribution of Galois orbits for mildly regular test functions

This paper establishes effective versions of Bilu's equidistribution theorem for Galois orbits of small-height points in the $N$-dimensional algebraic torus by developing a general Fourier analysis framework that quantifies convergence rates based on the regularity of mildly regular test functions, thereby extending prior results by Petsche and D'Andrea et al.

The Gist

Imagine you have a giant, multi-dimensional dance floor. This floor is shaped like a donut (or a collection of donuts, if we are in higher dimensions), which mathematicians call a polycircle. On this dance floor, there are thousands of dancers.

In the world of algebraic number theory, these dancers are special points called Galois orbits. Think of a Galois orbit as a group of "twins" or "cousins" of a single number. If you pick a complex number, its Galois orbit is the set of all its mathematical siblings that look different but share the same fundamental algebraic DNA.

The Big Idea: The "Bilu's Dance"

For a long time, mathematicians knew a beautiful fact about these dancers, proven by a man named Bilu in 1997.

The Rule: If you pick a sequence of these number-groups that are getting "smaller" and "simpler" (mathematically, their "Weil height" is approaching zero), they will eventually spread out perfectly evenly across the dance floor. They will stop clustering in corners and will form a uniform, smooth carpet of dancers covering the entire surface.

This is called Equidistribution. It's like pouring a bucket of glitter onto a table; eventually, the glitter settles into a perfect, even layer.

The Problem: "How Fast?" and "How Rough?"

While we knew that the glitter spreads out, we didn't know how fast it happens, especially if the glitter isn't perfectly smooth.

In the real world, we don't just look at perfect, smooth surfaces. We look at things with bumps, rough edges, or jagged shapes. In math, these are called test functions.

• Smooth functions are like silk sheets.
• Rough functions are like sandpaper or crumpled paper.

Previous studies (by researchers like Petsche and D'Andrea) could tell us how fast the dancers spread out, but only if we were looking at them through a smooth lens (like silk). They assumed the test functions were very "regular" (smooth, like Lipschitz continuous).

The Gap: What if we want to look at the dancers through a rough lens? What if the test function is only "mildly regular"—maybe it's a bit jagged, like a fractal or a function with a fractional derivative? The old math tools broke down here. There was a vast, unexplored world between "perfectly smooth" and "completely chaotic."

The Solution: A New Mathematical Microscope

This paper, by Emanuel Carneiro and Mithun Kumar Das, builds a new, powerful microscope using Fourier Analysis.

Think of Fourier Analysis as a way to break down a complex sound (or image) into its individual musical notes (frequencies).

• A smooth sound has mostly low, deep notes.
• A rough, jagged sound has a lot of high-pitched, sharp notes.

The authors realized that the "roughness" of the test function is directly related to how its "notes" (Fourier coefficients) behave. They developed a framework to measure exactly how the "roughness" of the lens affects the speed at which the dancers spread out.

The Key Findings (The "Recipe")

The paper provides a precise recipe that says:

"If your test function has a certain level of roughness (let's call it $\gamma$), then the error in the distribution (how uneven the dancers are) will shrink at a specific speed related to that roughness."

1. The Trade-off: The rougher your lens (the less regular the function), the slower the convergence. But the authors found the exact speed. It's not just "it gets better"; it's "it gets better at this specific rate."
2. The "Mildly Regular" Breakthrough: They proved that even if the function is only Hölder continuous (a fancy way of saying "somewhat smooth but with jagged edges"), the equidistribution still happens, and they can calculate exactly how fast.
3. The Limit: They found a natural limit to their method. You can't get infinite speed just by making the function smoother; there's a ceiling to how much the "roughness" can be tamed in this specific mathematical setup.

The Analogy of the "Rough Mirror"

Imagine you are trying to see a reflection of a crowd in a mirror.

• Old Math: You could only use a perfectly polished mirror. If the mirror was even slightly scratched, the old math couldn't tell you how blurry the reflection would be.
• This Paper: The authors figured out how to use scratched, frosted, and slightly cracked mirrors. They can now tell you: "If your mirror has scratches of this depth, the crowd will look this blurry, and here is the exact formula for that blurriness."

Why Does This Matter?

1. Precision: In science and engineering, we often deal with data that isn't perfectly smooth. This paper gives us the tools to handle "messy" data with mathematical precision.
2. New Bounds: In the appendices, they use these new tools to solve old problems about discrepancy (how unevenly points are distributed). They found better, tighter bounds for how points spread out in multi-dimensional spaces.
3. Completing the Puzzle: They used their new "rough lens" tools to provide a fresh, complete proof of the original theorem, showing that the theory holds up even under these new, more flexible conditions.

In a Nutshell

Carneiro and Das took a famous theorem about numbers spreading out evenly and asked, "What if we look at them with imperfect eyes?" They built a new mathematical framework using sound-wave analysis (Fourier) to measure exactly how "imperfect" our view can be and still get a reliable answer. They bridged the gap between "perfectly smooth" and "rough," giving us a sharper, more versatile tool for understanding the geometry of numbers.

Technical Summary

Here is a detailed technical summary of the paper "Effective Equidistribution of Galois Orbits for Mildly Regular Test Functions" by Emanuel Carneiro and Mithun Kumar Das.

1. Problem Statement

The paper addresses the problem of effective equidistribution for Galois orbits of sequences of algebraic points with small height in the $N$-dimensional algebraic torus $(\mathbb{Q}^\times)^N$.

• Context: Bilu's Equidistribution Theorem (1997) states that if $\{\xi_k\}$ is a "strict" sequence of points in $(\mathbb{Q}^\times)^N$ such that the Weil height $h(\xi_k) \to 0$, then the Galois orbits of these points become uniformly distributed on the unit polycircle $(S^1)^N$.
• The Gap: While Bilu's theorem guarantees convergence for any bounded continuous test function $F$, it does not provide a quantitative rate of convergence. Previous effective results (e.g., by Petsche, D'Andrea et al.) established bounds but required the test functions to be Lipschitz continuous.
• Objective: The authors aim to close the gap between continuous and Lipschitz functions by providing effective bounds for "mildly regular" test functions (specifically Hölder continuous functions and functions with fractional derivatives) using Fourier analysis. They seek to identify the precise quantitative dependence of the convergence error on the regularity of the test function and the height of the points.

2. Methodology

The paper employs a sophisticated Fourier analysis framework on the locally compact abelian group $(\mathbb{C}^\times)^N$, identified with $\mathbb{T}^N \times \mathbb{R}^N$ via logarithmic-polar coordinates $(\theta, s)$.

• Fourier Decomposition: The error term $E(F, \xi)$ (the difference between the average of $F$ over the Galois orbit and the integral over the uniform measure) is decomposed into two parts:
  1. Radial/Logarithmic part ($I_1$): Related to the deviation of the moduli $|\alpha|$ from 1.
  2. Angular part ($I_2$): Related to the distribution of the arguments $\arg(\alpha)$.
• Key Tools:
  • Siegel's Lemma (Bombieri-Vaaler version): Used to construct auxiliary polynomials with integer coefficients that vanish at the algebraic points with high multiplicity. This allows the authors to bound the sum of roots of unity.
  • Generalized Degree ($D(\xi)$): A crucial parameter defined as $D(\xi) := \min_{n \neq 0} \{ \|n\|_1 \deg(\chi_n(\xi)) \}$. This generalizes the standard degree of an algebraic number to the multidimensional setting and is shown to tend to infinity for strict sequences with vanishing height.
  • Tail Functions: The authors utilize tail functions of the Fourier transform (e.g., $\nu_{\hat{F}}(y)$) to handle cases where the test function does not have strong decay properties in the frequency domain.
  • Optimization of Parameters: The proofs involve optimizing a cutoff parameter (e.g., $M$ or $\delta$) to balance the contributions of low-frequency and high-frequency terms in the Fourier series.

3. Key Contributions and Results

The paper presents two main perspectives on regularity, leading to four primary theorems and their corollaries.

A. Regularity on the Fourier Space (Theorems 2 & 4)

This approach assumes conditions on the integrability of the Fourier transform $\hat{F}$.

• Theorem 2: Provides a bound involving non-decreasing functions $G$ and $H$. If $F$ has a Fourier transform with sufficient decay (weighted by $G$ and $H$), the error is bounded by terms involving $h(\xi)$ and $h_D(\xi)$.
  • Corollary 3 (Sharpness): For Hölder-type regularity (fractional derivatives of order $\gamma \le 1/2$), the error is bounded by $C(F) h_D(\xi)^\gamma$. The authors prove this exponent is sharp by constructing a specific counterexample where the error decays no faster than $h_D(\xi)^\gamma / |\log h_D(\xi)|$.
  • Improvement: This improves upon Petsche's (2005) result, which had an exponent of $1/3$for$N=1$, by achieving$1/2$ under milder regularity assumptions.
• Theorem 4: Handles cases where only the $L^1$ norm of the Fourier transform is finite, using tail functions to quantify the error.

B. Angular Regularity on Physical Space (Theorems 5 & 7)

This approach extends the work of D'Andrea, Narváez-Clauss, and Sombra (2017). It assumes:

1. Log-radial regularity: $F$ satisfies a modulus of continuity $\omega$ at $s=0$ (e.g., Hölder continuity in the radial direction).
2. Angular regularity: Conditions on the Fourier coefficients of the restriction $F_0(\theta) = F(\theta, 0)$.

• Theorem 5: Establishes a bound combining the radial modulus $\omega(2h(\xi))$ and a term depending on the angular Fourier coefficients weighted by a function $H$.
  • Corollary 6: For Hölder continuous functions (order $\gamma \le 1/2$) in the radial direction and fractional derivatives in the angular direction, the error is $O(h_D(\xi)^\gamma)$.
  • Significance: This yields the same quantitative estimate ($h_D(\xi)^{1/2}$) as the best previous Lipschitz results but requires weaker assumptions (Hölder-1/2 continuity instead of Lipschitz).
• Theorem 7: Provides a version for functions where only the angular part is in $L^1$, utilizing tail functions for the angular Fourier coefficients.

C. Applications and Extensions

• Appendix A (Angular Discrepancy): The authors apply their methods to bound the multidimensional angular discrepancy (where test functions are characteristic functions of intervals, which are discontinuous). They derive a bound of order $h_D(\xi)^{1/3} |\log h_D(\xi)|^{2(N-1)/3}$, refining previous results for $N=1$ and extending them to higher dimensions.
• Appendix C: A direct proof of Bilu's original theorem is provided using the effective estimates for a dense subclass of smooth functions, demonstrating the utility of the new bounds.

4. Significance

• Optimality: The paper establishes that the convergence rate of $h_D(\xi)^{1/2}$ is the natural limit for the Fourier analytic method used, even for functions with higher regularity. It proves that this rate is sharp for Hölder continuous functions.
• Bridging the Gap: By moving from Lipschitz to Hölder continuity and fractional derivatives, the authors significantly broaden the class of test functions for which effective equidistribution can be quantified.
• Multidimensional Generalization: The results generalize previous one-dimensional effective estimates (Petsche, Baker-Masser) to arbitrary dimensions $N \ge 1$ using a unified Fourier framework.
• Methodological Refinement: The introduction of the generalized degree $D(\xi)$ and the specific handling of the "tail" of the Fourier transform provide a robust toolkit for future problems in arithmetic geometry and equidistribution theory.

In summary, Carneiro and Das provide a comprehensive, sharp, and generalized theory for the effective equidistribution of Galois orbits, demonstrating that Fourier analysis can yield optimal convergence rates for test functions with significantly lower regularity than previously thought possible.