🔢 mathematics

Linear independence of values of hypergeometric functions and arithmetic Gevrey series

This paper establishes new linear independence results for the values of generalized hypergeometric functions at multiple distinct algebraic points over general number fields by introducing a uniform construction of Padé approximants and a novel non-vanishing argument for generalized Wronskians, thereby extending and strengthening previous findings in both complex and $p$ -adic settings.

Original authors: Sinnou David, Noriko Hirata-Kohno, Makoto Kawashima

Published 2026-07-07

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: Sinnou David, Noriko Hirata-Kohno, Makoto Kawashima

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ 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

Imagine you are a detective trying to solve a mystery involving a family of very special mathematical recipes. These recipes are called hypergeometric functions. They are like complex, infinite cooking instructions that, when followed, produce specific numbers.

The big question the authors of this paper are asking is: "If we take these recipes, mix them together with different ingredients (numbers), and add them up, can we ever get exactly zero?"

In the world of math, if the only way to get zero is to use "no ingredients" at all (meaning all the mixing coefficients are zero), then the recipes are considered linearly independent. If you can get zero by mixing them, they are dependent, and the math is less interesting. The authors want to prove that for a huge variety of these recipes, they are all unique and independent.

Here is how they did it, broken down into simple concepts:

1. The Three Types of Recipes

The authors realized these mathematical recipes come in three different "flavors" depending on how the ingredients are arranged (specifically, the relationship between two numbers, $p$ and $q$ ):

The "G-Function" Flavor ( $p = q + 1$ ): These are like well-behaved, infinite series that converge nicely. Think of them as recipes that work perfectly in both our normal world (real numbers) and in "p-adic" worlds (a strange, alternative way of measuring distance in math).
The "E-Function" Flavor ( $p < q + 1$ ): These are even more well-behaved, like the famous exponential function ( $e^x$ ). They are the "easy" ones that mathematicians have studied for a long time.
The "Z-Series" or "Euler" Flavor ( $p > q + 1$ ): These are the tricky ones. In our normal world, they explode and don't make sense (they have a radius of zero). However, in those alternative "p-adic" worlds, they actually work!

2. The Universal Tool: The "Magic Sieve"

To prove these recipes are independent, the authors needed a tool to test them. They built a universal "Magic Sieve" (mathematically called a Padé approximant).

The Analogy: Imagine you have a bunch of different soups (the functions). You want to know if they are all distinct. You take a sieve (the Padé approximant) and try to filter them.
The Innovation: Previous mathematicians had to build a different sieve for every single type of soup. The authors of this paper built one single, universal sieve that works for all three flavors of recipes, no matter how weird the ingredients are. This is a huge deal because it simplifies the whole process.

3. The "Non-Vanishing" Proof (The "Don't Disappear" Test)

The most critical part of their proof is showing that their "Magic Sieve" doesn't just disappear or turn into nothing (mathematically, the determinant doesn't vanish).

The Analogy: Imagine you are trying to prove that a group of people are all different. You ask them to stand in a specific formation. If the formation collapses into a single point, you can't tell them apart. But if the formation holds its shape and stays spread out, you know they are distinct.
The Breakthrough: The authors invented a new way to prove that their formation never collapses. They used a clever argument involving "generalized Wronskians" (a fancy way of measuring how spread out the functions are). They proved that no matter how you arrange the ingredients, the formation stays strong and distinct.

4. The Results: What Did They Find?

Using their universal sieve and their "don't disappear" proof, they confirmed:

For the "G-Function" and "E-Function" types: They proved that if you pick several different points (locations) to test these functions, the values you get are all independent. You can't mix them to get zero unless you use zero coefficients. They extended this from testing just one point to testing many points at once.
For the "Z-Series" (the tricky ones): They proved that in the alternative "p-adic" worlds, these values also don't have any hidden global relationships. They are independent.

5. Why Does This Matter?

The paper doesn't claim this will cure diseases or build bridges. Instead, it's a fundamental victory for Number Theory (the study of numbers).

The "Transcendence" Connection: In math, proving numbers are "linearly independent" is often the first step to proving they are "transcendental" (meaning they aren't the solution to any simple algebraic equation, like $\pi$ or $e$ ).
The "Universal" Aspect: By creating a method that works for all these different types of functions at once, the authors have given mathematicians a powerful, flexible tool. They showed that the "Magic Sieve" is robust enough to handle the easy cases, the hard cases, and the weird cases without needing to be rebuilt for each one.

In summary: The authors built a single, super-flexible mathematical tool that can prove that a vast family of complex number-generating recipes are all unique and independent, solving a problem that previously required many different, complicated tools. They did this by proving their tool never "collapses" into nothingness, ensuring the math holds up under scrutiny.

Technical Summary: Linear Independence of Values of Hypergeometric Functions and Arithmetic Gevrey Series

Problem Statement
The paper addresses the problem of establishing linear independence results for the values of generalized hypergeometric functions, denoted as ${}_pF_q$ , evaluated at several distinct algebraic points. The study operates within the framework of arithmetic Gevrey series, a class introduced by Y. André to extend the theory of Siegel's $E$ -functions and $G$ -functions. The authors aim to unify the treatment of three distinct parameter regimes:

$p = q + 1$ : The case of $G$ -functions (e.g., polylogarithms), where convergence is typically limited to the unit disk in the complex plane.
$p < q + 1$ : The case of $E$ -functions (e.g., the exponential series), which are entire functions.
$p > q + 1$ : The case of "Euler-type" or $Z$ -series, which have a radius of convergence of zero in Archimedean fields but positive radius in non-Archimedean ( $p$ -adic) fields.

Previous works often treated these cases separately or were restricted to single-point evaluations over the rational numbers. The paper seeks to provide a uniform construction for multi-point settings over general algebraic number fields, covering both Archimedean and non-Archimedean valuations.

Methodology
The core of the approach relies on the explicit construction of Padé approximants of type II for families of contiguous hypergeometric functions. The methodology proceeds through the following steps:

Operator Construction: The authors define a specific set of linear operators acting on the ring of polynomials $K[t]$ . Unlike previous methods that relied on derivation or primitivation (sufficient for polylogarithms), this work introduces an operator $T_c$ and a chain of differential operators linked to the parameters of the hypergeometric function. These operators allow for the systematic generation of a chain of functions linked by first-order differential operators.
Padé Approximation: Using these operators, the authors construct polynomials $P_\ell(z)$ and $P_{\ell,i,s}(z)$ that serve as Padé approximants for the hypergeometric functions evaluated at distinct points $\alpha_i$ . The construction is uniform across all parameter regimes ( $p < q+1$ , $p=q+1$ , $p>q+1$ ).
Non-Vanishing of the Wronskian: A critical component of the proof is establishing the non-vanishing of a generalized Wronskian of Hermite type (a determinant formed by the constructed approximants). The authors prove that this determinant is non-zero by analyzing the kernels of linear maps associated with the contiguous functions. They utilize a differential analogue of algebraic independence techniques, studying the intersection of kernels of specific morphisms to show that the only polynomial in the intersection is the zero polynomial.
Arithmetic Estimates: The paper provides detailed estimates for the norms of the constructed polynomials and the remainder terms of the Padé approximation in both Archimedean and non-Archimedean settings. These estimates are used to apply linear independence criteria (specifically, a qualitative criterion for $G$ -functions and a zero estimate for $E$ -functions and Euler-type series).

Key Contributions and Results

Uniform Construction: The paper presents a single, unified construction of Padé approximants that applies to all cases of $p$ and $q$ , removing the need for distinct methods for $E$ -functions, $G$ -functions, and Euler-type series.
Multi-Point Independence: The results extend known linear independence theorems from single-point to multi-point settings over general algebraic number fields.
- Case $p = q + 1$ : Theorem 2.3 establishes linear independence over $K$ for values of ${}_dF_{d-1}$ and their contiguous variants at distinct points $\alpha_i/\beta$ , provided a specific inequality involving heights and logarithmic terms ( $V_v(\alpha, \beta) > 0$ ) is satisfied. This generalizes results by Chudnovsky and Nesterenko.
- Case $p < q + 1$ : Theorem 2.4 proves linear independence over $\mathbb{Q}$ for values of ${}_pF_q$ and their contiguous variants at distinct algebraic points. This builds upon the Siegel-Shidlovsky theorem and recent work by Fischler, Rivoal, and Delaygue.
- Case $p > q + 1$ : Theorem 2.5 addresses the $p$ -adic setting for Euler-type series. It confirms that global linear relations do not exist among the values in the $p$ -adic setting, provided the points satisfy specific convergence conditions. The authors provide an effective bound on the size of the prime interval required to ensure non-vanishing of linear forms, which is shown to be a short interval proportional to $\log H / \log \log H$ .
Novel Non-Vanishing Proof: The paper introduces a new proof for the non-vanishing of the generalized Wronskian determinant. This proof is crucial for the universality of the method and relies on a detailed analysis of the kernels of differential operators rather than explicit computation of the determinant's value.
Quantitative Improvements: The authors claim that their construction allows for the replacement of technical analytic estimates with simpler norm operator evaluations, leading to better or competitive quantitative estimates.

Significance and Claims
The authors claim that their work generalizes and strengthens earlier results by providing a flexible framework for families of contiguous hypergeometric functions. The significance lies in:

Universality: The ability to treat $E$ -functions, $G$ -functions, and Euler-type series simultaneously under one Padé construction.
Extension to Multi-Points: Moving beyond single-point evaluations to handle systems of values at distinct algebraic points over general number fields.
Arithmetic Gevrey Series: The results are framed within the theory of arithmetic Gevrey series, offering new insights into the arithmetic properties of these formal solutions to differential equations.
Optimality in $p$ -adic Settings: For the case $p > q + 1$ , the paper provides a refined analysis of the prime intervals needed for non-vanishing, noting that their factorial estimates are optimal and do not suffer from the losses found in previous works (such as those by Matala-aho, Zudilin, or Chirskii).

The paper concludes by noting that while the method is powerful, it relies on the non-vanishing of the determinant, which is proven via kernel analysis rather than explicit calculation, and that the quantitative estimates are discussed in detail in Section 6. The work also includes a corrigendum to a previous lemma in their related literature regarding linear forms in polylogarithms.

1. The Three Types of Recipes

2. The Universal Tool: The "Magic Sieve"

3. The "Non-Vanishing" Proof (The "Don't Disappear" Test)

4. The Results: What Did They Find?

5. Why Does This Matter?

More like this