🔢 mathematics

Counting Frobenius extensions over local function fields

This paper determines the asymptotic growth of local function field extensions of characteristic $p$ with Galois groups contained in specific towers of cyclic extensions (including the affine group $AGL_1(p)$ , $S_3$ , and various Frobenius groups) when counted by discriminant.

Original authors: Jürgen Klüners, Raphael Müller

Published 2026-04-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jürgen Klüners, Raphael Müller

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 a master architect living in a magical city called Local Function Fields. This city is built on a strange, repeating grid system (characteristic $p$ ) where everything is made of tiny, repeating blocks.

Your job is to count how many different castles (mathematical extensions) you can build in this city. But there are rules:

The Blueprint: Every castle must follow a specific architectural style defined by a "Galois Group" (a set of symmetry rules).
The Budget: You can only build castles up to a certain "size" (measured by something called the discriminant).
The Goal: As your budget gets infinitely large, how does the number of possible castles grow? Does it grow like a gentle hill, a steep cliff, or a wild rollercoaster?

This paper by Jürgen Klüners and Raphael Müller is a detailed census of these castles, specifically for a very tricky type of architecture involving Frobenius Extensions.

Here is the breakdown of their adventure, translated into everyday language:

1. The Setting: The City of $F$

The city is a Local Function Field. Think of this as a place where numbers are like infinite strings of beads (Laurent series).

The Characteristic $p$ : The city has a weird rule: if you add a specific number to itself $p$ times, you get zero. This makes the math behave differently than in our normal world (like how a clock resets after 12 hours).
The Towers: The authors are looking at a specific construction method: a Tower of Towers.
- First, you build a base tower ( $L$ ) that is a "cyclic" extension (like a spiral staircase with $d$ steps).
- Then, you build a second layer ( $M$ ) on top of that base, which is a "cyclic" extension of order $p$ (like adding a small, $p$ -step spiral on top).

2. The Problem: Counting the Castles

Mathematicians have a famous guess (the Malle Conjecture) about how fast these castle counts grow. Usually, the formula looks like:
$\text{Count} \approx \text{Constant} \times (\text{Budget})^{\text{Exponent}}$

However, in this specific city (where the characteristic $p$ divides the size of the group), the usual rules break down. The "Exponent" in the formula changes, and the growth isn't smooth—it oscillates (wiggles up and down) like a heartbeat.

3. The Tools: Artin-Schreier Theory

To count these castles, the authors use a tool called Artin-Schreier Theory.

The Analogy: Imagine you have a giant bucket of water ( $L$ ). You want to find specific "ripples" in the water that create a new castle.
The tool tells you that every possible castle corresponds to a specific "ripple" (an element $\alpha$ ) in a vector space.
The "size" (discriminant) of the castle depends entirely on how "deep" or "high" the ripple is.

4. The Discovery: The "Oscillating Heartbeat"

The authors discovered two main things:

A. The Shape of the Growth

They found that the number of castles grows exponentially with the budget, but the rate depends on the shape of the group (the symmetry rules).

The Exponent: They calculated the exact "steepness" of the growth curve. It turns out the steepness depends on the degrees of the polynomial factors used to build the tower.
The Wiggle: The growth isn't a smooth line. It has a periodic function (a wiggly line) riding on top of the exponential growth.
- Metaphor: Imagine a rocket ship (the exponential growth) flying into space. But the rocket is shaking back and forth (the periodic function) as it flies. The shaking depends on the specific "residue" or remainder of the budget number.

B. The Two Ways to Count

The paper solves the problem in two different ways of looking at the castles:

Counting by Degree ($pd$): Looking at the castles as simple structures with $pd$ rooms.
- Result: The growth rate is determined by a specific fraction involving the group size.
Counting by Splitting Field ( $p^{\ell(I)}d$ ): Looking at the castles as fully symmetrical structures (Galois extensions).
- Result: The growth rate is slightly different, but still follows a predictable pattern with a wiggly heartbeat.

5. Why This Matters

Before this paper, we knew how to count castles in "nice" cities (where the group size and the city's characteristic don't clash). But in this "clashing" city (where $p$ divides the group size), the math was messy and unsolved for many complex shapes (like the $S_3$ group or Dihedral groups).

The authors' contribution:
They built a universal counting machine for these specific, tricky towers. They proved that even though the math is chaotic, the chaos follows a strict, predictable rhythm.

Summary Analogy

Imagine you are counting how many different ways you can arrange a stack of $p$ -colored bricks on top of a $d$ -colored base.

The Malle Conjecture is the rulebook for how the stack height grows as you get more bricks.
Usually, the rulebook says: "For every 100 bricks, you can make 10 new stacks."
But in this paper, the authors found that because the bricks are "sticky" (characteristic $p$ ), the rule changes.
They discovered that the number of stacks grows like $100^x$ , but the exact number wiggles up and down in a pattern that repeats every few thousand bricks.
They mapped out exactly how much it wiggles and how steep the growth is for every possible shape of the stack.

In a nutshell: They solved a complex counting puzzle for a specific type of mathematical structure, proving that even in the most chaotic algebraic environments, there is a hidden, rhythmic order to how things grow.

1. Problem Statement and Context

The Core Problem:
The paper addresses the asymptotic counting of field extensions $K/F$ over a local function field $F$ of characteristic $p$ , where the extensions are counted by their discriminant. Specifically, the authors focus on extensions where the Galois group (or the permutation group of the Galois closure) is a subgroup of the affine group $AGL_1(p) \cong C_p \rtimes C_d$ .

Context and Motivation:

Malle's Conjecture: For number fields and global function fields (where the characteristic is coprime to the group order), Malle's conjecture predicts the asymptotic growth of the number of extensions $Z(k, G; X)$ with discriminant bounded by $X$ as $X \to \infty$ . The formula is typically $c \cdot X^{a(G)} \log(X)^{b(G)-1}$ .
The Wild Case: When the characteristic $p$ divides the group order $|G|$ , the behavior changes significantly. Previous work (e.g., by Lagemann, Potthast) has shown that for elementary abelian $p$ -groups, the exponent $a(G)$ in Malle's conjecture is incorrect, and logarithmic factors may appear differently.
Gap: While abelian $p$ -groups and some specific non-abelian cases were studied, a general asymptotic formula for non-abelian groups in the "wild" ramification case (specifically Frobenius groups and subgroups of $AGL_1(p)$ ) over local function fields was lacking.

Specific Goal:
Determine the asymptotic behavior of the counting function $Z(F, G; X)$ for groups $G$ arising in towers $M/L/F$ where:

$L/F$ is a cyclic extension of degree $d$ (tame, $p \nmid d$ ).
$M/L$ is a cyclic extension of degree $p$ (wild, Artin-Schreier).
The resulting Galois group $G = \text{Gal}(M/F)$ is a subgroup of $AGL_1(p)$ .

2. Methodology

The authors employ a combination of Artin-Schreier theory, representation theory of finite groups, and analytic number theory (Dirichlet series).

A. Structural Setup

Field Tower: Let $F = \mathbb{F}_q((t))$ be a local function field. Let $L/F$ be a cyclic extension of degree $d$ ( $p \nmid d$ ).
Artin-Schreier Extensions: Extensions $M/L$ of degree $p$ are classified by the cokernel of the Artin-Schreier operator $\wp(x) = x^p - x$ . The space $L/\wp(L)$ is an $\mathbb{F}_p$ -vector space.
Module Structure: The Galois group $H = \text{Gal}(L/F) \cong C_d$ acts on $L/\wp(L)$ , making it an $\mathbb{F}_p[H]$ -module. The authors decompose this module based on the factorization of $X^d - 1$ over $\mathbb{F}_p$ .

B. Classification of Groups

The polynomial $X^d - 1$ factors into irreducibles $f_1(X) \cdots f_r(X)$ over $\mathbb{F}_p$ .

The authors define a set of indices $I \subseteq \{1, \dots, r\}$ .
They construct specific subgroups $G_p(d, I) \cong M_I \rtimes C_d$ , where $M_I$ is a direct sum of cyclic modules corresponding to the factors in $I$ .
These groups cover various non-abelian cases, including $S_3$ (for $p=2, d=3$ ), dihedral groups $D_{2p}$ , and general Frobenius groups $C_p \rtimes C_d$ .

C. Discriminant Analysis

Generators: An extension is generated by an element $\alpha \in L/\wp(L)$ . The Galois group of the splitting field depends on the $\mathbb{F}_p[H]$ -submodule generated by $\alpha$ .
Discriminant Formula: The discriminant exponent is derived using the conductor of the Artin-Schreier extension. For a generator $\alpha$ , the discriminant depends on the valuation $\nu_L(\alpha)$ and the specific submodule structure.
Tower Formula: The authors use the discriminant tower formula to relate $\text{disc}(M/F)$ to $\text{disc}(M/L)$ and $\text{disc}(L/F)$ .

D. Counting Strategy

Decomposition: The counting problem is reduced to counting elements $\alpha$ in specific subspaces $Y_I \subset L/\wp(L)$ with bounded valuation.
Periodicity: The structure of the module $L/\wp(L)$ is periodic with respect to the valuation. This allows the counting function to be expressed as a sum over arithmetic progressions.
Dirichlet Series: For the Galois case (counting splitting fields), the authors construct a Dirichlet series $\Phi(s) = \sum q^{-\text{disc}(M/F)s}$ .
Analytic Continuation: They decompose this series into sub-sums based on the "composition" of valuations of the components of $\alpha$ . Using techniques similar to those in Potthast's work on elementary abelian groups, they analyze the poles of these series.
Tauberian Theorems: The asymptotic growth is determined by the location of the rightmost pole of the Dirichlet series.

3. Key Contributions and Results

The paper provides explicit asymptotic formulas for two distinct counting scenarios:

Result 1: Counting Fields of Degree $pd$ (Non-Normal Extensions)

The authors count extensions $K/F$ of degree $pd$ where the Galois group of the Galois closure is $G_p(d, I)$ , viewed as a permutation group on $pd$ points.

Theorem 4.5: There exists a periodic function $\beta_I(x)$ $β_{I} (x)$ such that:
$Z_{pd}(F, G_p(d, I); x) \sim \beta_I(x) q^{a x}$
where the exponent is $a = \frac{\ell(I)}{pd}$ $a = \frac{ℓ ( I )}{p d}$ .
- Here, $\ell(I) = \sum_{i \in I} \deg(f_i)$ .
- The function $\beta_I(x)$ is $(p-1)pd$-periodic.
- Significance: The asymptotic growth depends only on the order of the group ($pd$) and the specific subgroup structure encoded in $\ell(I)$ . The oscillating term $\beta_I(x)$ captures the finer arithmetic structure.

Result 2: Counting Galois Extensions (Splitting Fields)

The authors count the actual Galois extensions $M/F$ (splitting fields) where the Galois group is $G_p(d, I)$ , viewed as a permutation group on $p^{\ell(I)}d$ points.

Theorem 5.1: There exists a periodic function $\tilde{\beta}_I(x)$ $\tilde{β}_{I} (x)$ such that:
$Z_{p^{\ell(I)}d}(F, G_p(d, I); x) \sim \tilde{\beta}_I(x) q^{\tilde{a} x}$
where the exponent is:
$\tilde{a} = \frac{p-1}{pd} \cdot \frac{\ell(I)}{p^{\ell(I)} - 1}$
- Significance: This exponent $\tilde{a}$ differs from the naive Malle prediction. It reflects the "wild" nature of the extension. The term $\frac{p-1}{p^{\ell(I)}-1}$ arises from the specific geometry of the Artin-Schreier tower and the discriminant formula for elementary abelian $p$ -extensions.
- The periodicity is $(p^{\ell(I)} - 1)pd$ .

Specific Cases

The results apply to:

Dihedral Groups $D_{2p}$ : When $\ell(I)=1$ and $d=2$ (for $p>2$ ).
Frobenius Groups $C_p \rtimes C_d$ : General cases where $d | (p-1)$ .
$S_3$ : For $p=2, d=3$ .
$A_4$ and $C_2^3 \rtimes C_7$ : Specific cases for $p=2$ .

4. Significance and Implications

Resolution of the Wild Case: The paper successfully extends the understanding of discriminant distribution to non-abelian groups in the wild ramification setting ( $p \mid |G|$ ), a notoriously difficult area in arithmetic statistics.
Refutation/Correction of Malle's Conjecture: The results confirm that for groups where $p \mid |G|$ , the exponent $a(G)$ in Malle's conjecture is not simply determined by the minimal index of a conjugacy class in the same way as in the tame case. The derived exponents $\tilde{a}$ are specific to the structure of the $p$ -Sylow subgroup and its interaction with the tame part.
Oscillatory Behavior: The paper highlights that the counting functions are not smooth power laws but involve periodic oscillating functions ( $\beta_I(x)$ ). This is a characteristic feature of counting problems over local fields with wild ramification, reflecting the discrete valuation structure.
Methodological Advancement: The authors develop a robust framework for handling the discriminant of splitting fields of non-abelian extensions by reducing the problem to the study of $\mathbb{F}_p[H]$ -modules and analyzing the poles of associated Dirichlet series. This methodology can likely be adapted to other non-abelian $p$ -groups.
Unification: The work unifies the counting of various Frobenius groups under a single theoretical umbrella, showing that their asymptotic behavior is governed by the same structural parameters ( $\ell(I)$ and the group order).

In summary, Klüners and Müller provide a complete and rigorous solution to the counting problem for a broad class of non-abelian extensions over local function fields, revealing complex periodic behaviors and correcting previous heuristic expectations for the wild case.

1. The Setting: The City of FFF