🔢 mathematics

Galois-invariant Néron--Severi ranks of Fermat surfaces over number fields: a Galois module, closed forms, a threshold, and exact tables

This paper provides an explicit, unconditional computation of the Galois-invariant Neron-Severi ranks for Fermat surfaces over arbitrary number fields by deriving closed-form character averages for degrees coprime to six, establishing field-of-definition thresholds, and presenting exact rational rank tables for degrees up to 30.

Original authors: Rifat Jumagulov

Published 2026-07-21

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Rifat Jumagulov

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

Imagine you are an architect trying to understand the hidden skeleton of a building made of pure light. In the world of mathematics, specifically a branch called algebraic geometry, these "buildings" are shapes called surfaces. Just like a physical building has beams, pillars, and walls that hold it together, these mathematical surfaces have invisible structural lines and planes that define their shape. Mathematicians call this hidden framework the "Néron–Severi group." It's like counting how many independent ways you can draw a line or a plane on the surface without it falling apart or merging into something else.

Now, imagine these surfaces aren't just sitting in a quiet room; they are being watched by a mischievous group of "gardeners" called Galois groups. These gardeners don't just look at the building; they rearrange the light, swap the colors, and twist the angles according to strict rules. Sometimes, a structural beam that looks solid from one angle disappears or changes when the gardeners twist the view. The big question mathematicians have been asking is: "How many of these structural beams stay exactly the same, no matter how the gardeners twist the light?" This number is called the "Galois-invariant rank." It tells us how much of the surface's structure is truly "real" and permanent, rather than just an illusion created by a specific viewpoint.

This paper, written by Rifat Jumagulov, tackles this question for a very special family of surfaces called "Fermat surfaces." Think of these as the most symmetrical, perfect-looking surfaces you can imagine, defined by a simple equation where four variables are added together, each raised to the same power (like $x^4 + y^4 + z^4 + w^4 = 0$ ). The author has cracked the code to count exactly how many of these permanent structural beams exist for these surfaces, providing precise formulas for many cases and a complete, exact table for every power from 4 up to 30.

The main discovery is a set of precise formulas and tables that act like a master key. For many of these surfaces—specifically when the power is not divisible by 2 or 3—the author found a simple, closed-form recipe to calculate the number of permanent beams. For example, if the power is a prime number like 5, 7, or 11 (as long as it's not divisible by 2 or 3), the number of permanent beams follows a neat pattern: it's roughly three times the power, minus five. So, for a surface with power 5, there are exactly 10 permanent beams; for power 7, there are 16. The author didn't just guess this; they proved it using deep tools from number theory, showing that for these specific numbers, the Galois action removes almost all of the geometric rank, leaving a much smaller, specific number of beams that survive the twist.

However, the story gets trickier when the power is an even number or divisible by 3. In these cases, the gardeners are more aggressive, and some beams that looked solid from a distance turn out to be illusions. The paper reveals a "threshold": if the power is divisible by 3, the gardeners break the symmetry, and the number of permanent beams drops below the maximum possible. The author spent a lot of time figuring out exactly how many beams survive in these messy cases. They created a detailed table for powers ranging from 4 up to 30, listing the exact count for each. Some of these counts were surprisingly tricky to find, requiring the author to use special mathematical identities (like the Hasse–Davenport identities) to prove that certain beams were indeed permanent, while others were not.

The paper also explains why the gardeners behave this way. It turns out that the "twist" the gardeners apply depends on the specific power of the surface. For some powers, the twist is trivial (meaning the gardeners do nothing), and for others, it's a complex rotation that cancels out certain beams. The author describes the entire collection of beams as a "monomial Galois module," which is a fancy way of saying they mapped out exactly how the gardeners shuffle the beams around. They showed that for any field of numbers you choose, you can calculate the number of permanent beams by looking at how these gardeners interact with that specific field.

One of the most exciting parts of the paper is the "average" result. If you were to pick a random power (that isn't divisible by 2 or 3) and count the beams, the author proved that on average, the number of beams grows in a predictable way, roughly proportional to the square of the power. This gives a big-picture view of how these surfaces behave as they get more complex.

The author is incredibly confident in these results. They didn't just simulate the numbers; they provided rigorous proofs for every single entry in their tables, covering all degrees from 4 to 30. For the tricky cases where the math got complicated (like powers 14, 24, 28, and 30), they used a combination of advanced algebraic tricks and computer-assisted verification to ensure every count was exact. They even made their code and data available for anyone to check, proving that their "master key" works perfectly. In short, this paper takes a complex, twisting problem about invisible mathematical structures and turns it into a clear, exact map, showing us exactly how many beams hold up these beautiful, symmetrical surfaces, no matter how the gardeners try to twist the light.

Technical Summary: Galois-Invariant N´eron–Severi Ranks of Fermat Surfaces Over Number Fields

Problem Statement
The paper addresses the computation of the Galois-invariant N´eron–Severi rank, denoted $\rho_K(X_d)$ , for Fermat surfaces $X_d$ defined by the equation $x_0^d + x_1^d + x_2^d + x_3^d = 0$ over an arbitrary number field $K$ . While the geometric Picard number $\rho_{\mathbb{Q}}(X_d)$ (the rank over the algebraic closure) is classical and well-understood via the work of Shioda, Aoki, and others, the rank over specific number fields $K$ —specifically the dimension of the Galois invariants $\dim((NS(X_{\mathbb{Q}}) \otimes \mathbb{Q})^{\text{Gal}(\overline{\mathbb{Q}}/K)})$ —lacked a unified, explicit treatment for the Fermat family. The challenge lies in determining which algebraic classes (specifically the $(1,1)$ -classes identified by the Lefschetz theorem) are defined over $K$ , a problem complicated by the action of the Galois group on the eigenspaces of the cohomology and the presence of non-trivial twists (Gr¨ossencharacters) and stabilizer cocycles.

Methodology
The author employs a character-theoretic approach grounded in the eigenspace decomposition of the primitive cohomology $H^2_{\text{prim}}(X_d, \mathbb{C})$ . The methodology proceeds through several key steps:

Monomial Module Formalism: The N´eron–Severi group is treated as a monomial Galois module. The primitive part is decomposed into eigenspaces $V(a)$ indexed by algebraic characters $a = (a_1, \dots, a_4)$ . The Galois action permutes these lines and acts on them via scalars (twists and stabilizer characters).
Reduction to Field of Definition: The paper establishes a reduction $\rho_K(X_d) = \rho_{K \cap L}(X_d)$ , where $L$ is the field of definition of the geometric Picard group (determined by Gvirtz-Chen–Skorobogatov). This reduces the problem to computing invariants over subfields of $L$ .
Orbit Counting with Twists: Using Frobenius reciprocity, the rank is expressed as a sum over orbits of characters under the action of $(\mathbb{Z}/d)^\times$ . An orbit contributes 1 to the rank if and only if both its vertical twist (Gr¨ossencharacter) and its stabilizer cocycle character are trivial.
Explicit Evaluation of Twists: The vertical twists are evaluated using Gauss sums and Hasse–Davenport product relations. The author distinguishes between:
- Decomposable characters: Those that can be paired into zero-sum pairs.
- Family characters: Those belonging to specific one-parameter families (Aoki–Shioda families).
- Exceptional characters: Those not falling into the above categories, occurring at specific "exceptional levels."
Stabilizer Descent: For even degrees, a "sub-level reduction" is used to evaluate the stabilizer cocycle. This involves lifting the Galois action to a sub-level character and checking parity conditions on the pair-representative sums.

Key Contributions and Results

Closed Formulas for Coprime Degrees: For degrees $d$ coprime to 6, the paper derives a closed-form formula for the rank over $\mathbb{Q}$ :
$\rho_{\mathbb{Q}}(X_d) = 1 + 3(\Psi_2(d) - 3\tau(d) + 2)$
where $\Psi_2$ is a multiplicative function and $\tau$ is the divisor count. Specifically, for prime $p \geq 5$ , $\rho_{\mathbb{Q}}(X_p) = 3p - 5$ .
Field-of-Definition Threshold: The paper proves a dichotomy: if $\gcd(d, 6) = 1$ , every algebraic character is decomposable and has a trivial twist, meaning the rank equals the "orbit ceiling" (the maximum possible rank). If $\gcd(d, 6) > 1$ , non-trivial twists exist (specifically Kummer characters associated with radicals like $2^{1/(d/2)}$ or $3^{1/(d/3)}$ ), causing the rank to drop below the ceiling.
Exact Tables for $4 \leq d \leq 30$ : The author provides exact values for $\rho_{\mathbb{Q}}(X_d)$ $ρ_{Q} (X_{d})$ for all degrees up to 30. This includes handling exceptional cases at $d = 14, 24, 28, 30$ $d = 14, 24, 28, 30$ where non-decomposable orbits with trivial vertical twists contribute to the rank. These contributions ( $E_d$ $E_{d}$ ) are proven using "doubling telescopes," sub-level reductions, and direct Hasse–Davenport identities.
- Examples: $\rho_{\mathbb{Q}}(X_4) = 5$ , $\rho_{\mathbb{Q}}(X_{14}) = 46$ , $\rho_{\mathbb{Q}}(X_{24}) = 98$ .
General Number Fields: The paper provides a general reduction formula for any number field $K$ , expressing $\rho_K$ as an average of the character of the Galois action over the subgroup fixing $K \cap L$ . This allows for the computation of ranks over cyclotomic subfields and radical extensions.
Asymptotic Average Order: The paper establishes the asymptotic behavior of the rank for degrees coprime to 6:
$\sum_{d \leq x, \gcd(d,6)=1} \rho_{\mathbb{Q}}(X_d) \sim \frac{3}{5}x^2$
indicating that the average rank grows linearly with $d$ (specifically $\sim \frac{18}{5}d$ ).

Significance and Claims
The paper claims to provide the first closed-form formulas and comprehensive exact tables for the Galois-invariant N´eron–Severi rank of the Fermat family over $\mathbb{Q}$ . Unlike prior work which focused on geometric ranks or specific small degrees (like $d=4$ ), this work unifies the treatment of the rank over arbitrary number fields via a character-theoretic framework.

The author emphasizes that all results are unconditional, relying on the Lefschetz (1,1) theorem to identify algebraic classes with rational $(1,1)$ -classes. The paper distinguishes itself by:

Explicitly describing the N´eron–Severi group as a monomial Galois module.
Evaluating the rank as a character average, providing closed forms where possible and exact tables where necessary.
Resolving the "exceptional" cases ( $d=14, 24, 28, 30$ ) through rigorous internal identities (telescopes and stabilizer descents) rather than relying solely on computational sampling.
Providing machine-readable certificates and standalone scripts that verify every load-bearing computation, including exact arithmetic in cyclotomic fields and point-counting checks.

The work is presented as a definitive calculation of these invariants for the Fermat family, reducing the problem of finding $\rho_K$ to a formal representation-theoretic computation once the field of definition $L$ is known.

More like this