On the Weil descent of Artin-Schreier algebraic function fields over finite fields

Robert Rolland; St\'ephane Ballet

REVIEW 2 major objections 4 minor 14 references

On the Weil descent of Artin-Schreier algebraic function fields over finite fields

T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read For generalized Artin-Schreier extensions over $F_{p^4}$, the paper proves that descent to $F_p$ is exactly controlled by the Frobenius action on the additive Galois group, yielding a complete count of which degree-$p$ subextensions…

desk verdict Useful explicit classification of bi-cyclic Artin-Schreier descents, with a false general intersection lemma that needs correction. read the letter →

arxiv 2505.21656 v1 pith:4AWI57DO submitted 2025-05-27 math.NT

Stéphane Ballet , Robert Rolland This is my paper

classification math.NT MSC 11G2014H25

keywords algebraicfunctionfieldsfinitedescentofdefinitionfieldWeilArtin-SchreierextensionsGaloisgroupFrobeniusactionbi-cyclic

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks when a generalized Artin-Schreier function field $F$ over $K=\mathbb{F}_{p^{2s}}(x)$, defined by $y^{p^s}+y=f(x)$, can have its definition field lowered to a smaller finite field $\mathbb{F}_{p^t}$ with $t$ dividing $2s$. The main general result is a necessary and sufficient criterion: a subextension $E$ of $F$ descends to $\mathbb{F}_{p^t}$ if and only if the subgroup $H$ of the additive Galois group $G$ fixing $E$ is invariant under the Frobenius action. In the bi-cyclic case $s=2$, where $G$ is a two-dimensional $\mathbb{F}_p$-vector space, this criterion gives a complete classification of the degree-$p$ subextensions: exactly two descend to $\mathbb{F}_p$ when $p\equiv 1\pmod 4$, none when $p\equiv 3\pmod 4$, and exactly one when $p=2$, while every such subextension descends to $\mathbb{F}_{p^2}$. This matters because descent of the definition field is the algebraic translation of asking whether a curve over a large finite field admits a model over a smaller field.

What carries the argument

The central object is the additive Galois group $G=\{\alpha\in K:\alpha^{p^s}+\alpha=0\}$, whose elements are the automorphisms $f_\alpha:y\mapsto y+\alpha$; it is an $\mathbb{F}_p$-vector space of dimension $s$. The descent criterion (Theorem 3.4) says that a subextension descends exactly when its stabilizer subgroup $H$ is stable under the action $\Theta(\sigma)(f_\alpha)=f_{\sigma(\alpha)}$, which in the finite-field case is just the action of Frobenius on $G$. In the bi-cyclic case $s=2$, $G$ is a plane over $\mathbb{F}_p$ with $p+1$ one-dimensional subspaces, and the classification reduces to counting which lines are preserved by the Frobenius map $\alpha\mapsto\alpha^p$.

What would settle it

Take $p=3$, $K=\mathbb{F}_{81}$, and any $f(x)$ satisfying the paper's conditions, so $F$ is defined by $y^9+y=f(x)$. The paper predicts that the four degree-$3$ subextensions of $F/\mathbb{F}_{81}(x)$ come in two Frobenius-swapped pairs and none has a model over $\mathbb{F}_3$. Finding a model over $\mathbb{F}_3$ for any one of them, or equivalently finding a nonzero $a\in\mathbb{F}_{81}$ with $a^9+a=0$ such that $\langle a\rangle$ is stable under $a\mapsto a^3$, would refute the classification.

Watch

Extended reading notes

Core claim

In the language of the paper: let $K=\mathbb{F}_{p^{2s}}$, let $F/K(x)$ be the generalized Artin-Schreier extension with additive Galois group $G=\mathrm{Gal}(F/K(x))\cong\{\alpha\in K:\alpha^{p^s}+\alpha=0\}$, and let $\Gamma=\mathrm{Gal}(K(x)/k(x))$. Then the fixed field $E$ of a subgroup $H\subseteq G$ can be reduced over $k$ if and only if $H$ is globally invariant under $\Theta(\sigma)$ for every $\sigma\in\Gamma$, where $\Theta(\sigma)$ sends $f_\alpha$ to $f_{\sigma(\alpha)}$ (Theorem 3.4). For $s=2$, $K=\mathbb{F}_{p^4}$ and $G$ has $p+1$ subgroups of order $p$. The Frobenius $\varphi_2$ acts on $G$ as multiplication by $-1$ (and as the identity when $p=2$), so every one-dimensional subgroup is stable and every degree-$p$ subextension descends to $\mathbb{F}_{p^2}$. The Frobenius $\varphi_1$ acts on the $p+1$ subgroups as a permutation: when $p\equiv 1\pmod 4$ it fixes exactly two, when $p\equiv 3\pmod 4$ none, and when $p=2$ exactly one, giving the stated classification.

Load-bearing premise

The classification rests on the classical theorem that stability of a subgroup under the Galois group $\Gamma$ is sufficient for its fixed field to descend, and this sufficiency is only guaranteed when $F/k(x)$ is Galois with group the semidirect product $G\rtimes\Gamma$; the paper secures that by assuming the defining polynomial $A(x)$ has coefficients in $k[x]$. If that condition is dropped, the 'if' direction of the criterion could fail and the bi-cyclic counts might be incomplete.

Editorial extensions

If this is right

Every degree-$p$ subextension of $F/\mathbb{F}_{p^4}(x)$ descends to $\mathbb{F}_{p^2}$, because $\varphi_2$ sends each element of $G$ to its negative and therefore preserves every subgroup.
If $p\equiv 1\pmod 4$, exactly two of the $p+1$ degree-$p$ subextensions have models over $\mathbb{F}_p$; the paper gives their generators explicitly from a primitive element of $\mathbb{F}_{p^4}$.
If $p\equiv 3\pmod 4$, none of the degree-$p$ subextensions descends to $\mathbb{F}_p$; the Frobenius $\varphi_1$ swaps them in pairs.
For $p=2$, there are three degree-$2$ subextensions, all descending to $\mathbb{F}_4$, and exactly one of them descends to $\mathbb{F}_2$.
The general criterion turns descent into a finite linear-algebra check: test stability of $H$ under the Frobenius action instead of searching for a rational model.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

One consequence the paper does not spell out is that for any intermediate $k=\mathbb{F}_{p^t}$, the number of degree-$p$ subextensions descending to $k$ should be the number of one-dimensional $\mathbb{F}_p$-subspaces of $G$ stable under $\alpha\mapsto\alpha^{p^t}$; for $t=s$ this is all of them, since the map is $-\mathrm{Id}$ (or Id when $p=2$).
The count of descending subextensions depends only on the Frobenius action on $G$, so changing the rational function $f(x)$ within the allowed class should not change the classification, only the actual fields and models.
The same stability criterion could be applied to $s>2$ by counting Frobenius-fixed subspaces of $G\cong(\mathbb{Z}/p\mathbb{Z})^s$; the paper does not attempt this full classification, but the linear-algebra translation is already in place.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Dear colleague,

The short version: the s=2 classification of descendable subextensions is real, correct, and useful, but the paper overclaims a general lemma that is false. If you take the bi-cyclic results (Theorems 4.6-4.8) and ignore Theorem 4.4, you're fine; the paper needs work before the general part can stand.

What's actually new: for A(T)=T^{p^2}+T over F_{p^4}, the paper computes which order-p subgroups of the Galois group are Frobenius-stable. For odd p, it finds exactly two stable subgroups when p≡1 mod 4, none when p≡3 mod 4; for p=2 exactly one, and all subgroups descend to F_{p^2}. That's a clean, concrete result that I don't think is in the prior literature. The descent criterion itself is recalled from Weil/Serre and the authors' earlier paper, but the stability computation for this family is the new piece. The explicit examples for p=5,13,17 are reproducible and help.

The soft spots are not in the s=2 computation itself. Theorem 4.4 claims k_t ∩ G = {0} for all t | 2s. That is false. For t=2s, k_t=K and the intersection is all of G; for p=2, the statement fails already at t=1 because the Frobenius condition a^{p^s}+a=0 becomes a^{p^s}=a, so G=F_{p^s} and k_1=F_p sits inside. The proof also uses (φ_t)^{s/t}, which is not an integer power when t ∤ s. This doesn't harm the s=2 classification—those proofs don't invoke 4.4—but the abstract promises general results for all t dividing 2s, and that generality is unsupported. The clean fix is to restrict the claim to odd p and to t dividing s, or to state the p=2 situation separately. There's also a minor gap in Theorem 4.7: it exhibits two stable subgroups and asserts uniqueness without a rigorous counting argument. For a paper of this size, that's a small repair, but it should be added.

Net: the paper is worth a serious referee. The core classification is correct, the exposition is mostly clear, and the errors are localized and fixable. I'd recommend conditional acceptance with a request to correct Theorem 4.4 and tighten Theorem 4.7. Someone working on function fields with many points or algebraic geometry codes will find the classification useful; for the rest of us, it's a solid but narrow contribution.

Referee Report

2 major / 4 minor

Summary. The paper studies the descent of the definition field for generalized Artin-Schreier extensions F/F_{p^{2s}}(x) defined by y^{p^s}+y=f(x), where K=F_{p^{2s}} is the full constant field. It states a general criterion (Theorem 3.4) asserting that a subextension E corresponding to a subgroup H⊆G=Gal(F/K(x)) descends to k=F_{p^t} if and only if H is stable under the Frobenius action. The paper then specializes to s=2 and classifies the subextensions descending to F_p and F_{p^2}: for odd p, exactly two degree-p subextensions descend to F_p when p≡1 mod 4, and none when p≡3 mod 4; for p=2, exactly one descends to F_2 and all three descend to F_4. The appendix lists explicit invariant subgroups for p=5, 13, and 17.

Significance. If the classification were correct, it would give a complete, explicit answer to a natural descent question for Artin-Schreier function fields with constant field F_{p^4}, with potential applications to towers of function fields and to the theory of definition-field descent. The paper's computational examples are explicit and checkable, and the stability criterion is a clean formulation. However, the central criterion is false as stated, and a concrete counterexample in the paper's main p=2, s=2 case invalidates Corollary 4.9. The general statement of Theorem 4.4 is also false. These are load-bearing errors, not presentation issues, so the paper cannot be accepted in its current form.

major comments (2)

[Section 3.2, Theorem 3.4] The 'if' direction of Theorem 3.4 is false as stated. Take p=2, s=2, k=F_2, K=F_16, and f=α x^5 with α a primitive element of F_16. This f satisfies the standing conditions: the pole at infinity has odd order 5. The subgroup G_1={0,1} of G={a∈F_16 : a^4+a=0} is invariant under φ_1. The corresponding fixed field is E_1=K(x,z) with z^2+z=αx^5. If E_1 could be reduced to F_2, then Theorem 2.3 would force E_1/F_2(x) to be Galois. But the conjugate equation T^2+T=α^2x^5 has no root in E_1: a root w would give u=w+z∈E_1 with u^2+u=x^5, and writing u=a+bz with a,b∈F_16(x) forces b∈{0,1} and then a^2+a=x^5 or a^2+a=(1+α)x^5; both are impossible because a^2+a has even pole order at infinity while the right-hand side has pole order 5. Hence E_1 does not descend to F_2, contradicting both Theorem 3.4 and Corollary 4.9's claim that the F_2-invariant subextension is the unique one descending to F_2.
[Section 4.1, Theorem 4.4] Theorem 4.4 is false as stated. The proof applies (φ_t)^{s/t}, which is only an integer power when t divides s, but the theorem is announced for every t dividing 2s. For t=2s, k_t=K and G is a nonzero subgroup of K, so k_t∩G=G≠{0}. For p=2 and t|s, every u∈k_t satisfies u^{p^s}=u, hence u^{p^s}+u=0, so k_t⊂G; the s=2 example of §4.3.3 already has G=F_4 and k_1=F_2⊂G. The statement needs to be restricted to odd p and t|s, or replaced by a correct computation of k_t∩G. Because the abstract promises general results for all t dividing 2s, this is a load-bearing error.

minor comments (4)

[Section 4.3.1, proof of Theorem 4.6] There is a typo 'il' for 'if', and the non-existence of a square root of -1 in F_p when p−1 is not divisible by 4 is usually justified by Euler's criterion rather than Wilson's theorem; the reference to Wilson's theorem is imprecise.
[Theorem 4.2] For t=2s, the displayed set contains the duplicate β^{p^{2s}-1}=1 and does not list the correct last exponent p^{2s}-2; the formula should list exponents 0 through p^t−2.
[Section 4.3.3] The text refers to 'the results of Section 3.5', but no Section 3.5 exists; the intended reference is likely Section 3.2 or Section 3.4.
[Remark 4.10] The additive polynomial is written T^{p^s}+T^p, which conflicts with the paper's standing polynomial T^{p^s}+T; if this is intentional, it needs an explanation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the descent criterion is quoted from classical Weil/Serre descent theory, and the bi-cyclic Frobenius-stability classification is computed directly from the definition of G.

full rationale

The paper's central claim (Theorems 3.4, 4.6, 4.7, 4.8 and Corollary 4.9) does not derive its target classification from assumptions that already contain it. Theorem 3.4 combines Proposition 3.3 with Theorem 2.2, quoted from the authors' earlier paper [2] but explicitly traced to Weil [14] and Serre [11], and with Theorem 2.3. The sufficiency direction of the descent criterion is a standard Galois-descent result, not a restatement of the conclusion; moreover the paper verifies the semidirect-product hypothesis by assuming A(x) in k[x], which makes F/k(x) Galois with group G semidirect Gamma, so the criterion is not being invoked vacuously. The finite-field content — the elements of G, Frobenius orbits on the subgroups G_i, and the p=2 stabilization check — is computed directly from the definition G = {alpha in F_{p^{2s}} : alpha^{p^s} + alpha = 0}, and the asserted stable subgroups are exhibited as solutions of explicit equations. The one self-citation to [2] for Theorem 2.2 is not load-bearing in a circular way: it supplies a classical descent criterion whose hypotheses are checked in the present setting, and the bi-cyclic classification does not reduce to any fitted parameter or to the conclusion of [2]. The apparent error in Theorem 4.4 (it appears to fail when t=2s, and also for p=2 when t=1) is a correctness and internal-consistency issue, not a circularity issue: that theorem is not used in the explicit s=2 classification, and its failure does not make any 'prediction' equivalent to an input by construction. No fitted input is renamed as a prediction, no uniqueness theorem is imported from the authors alone, and no ansatz is smuggled in through self-citation. The classification is therefore self-contained apart from the standard descent background, and no circular step can be exhibited.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claims rest on standard descent theory, a standard characterization of Artin-Schreier extensions, and finite-field facts. No free parameters are fitted and no new entities are postulated.

assumptions (4)

domain assumption Weil's descent theorem: if F/k(x) is Galois with group G semidirect Gamma, then the descent from K to k is possible (Theorem 2.2).
Invoked via Theorem 2.2 quoted from ref. [2], [11], and [14]. The paper relies on this for the 'if' direction of Theorem 3.4.
standard math Stichtenoth's characterization of Artin-Schreier extensions (Proposition 3.1): the Galois group G is isomorphic to the additive subgroup of roots of the linearized polynomial A(T).
Used to identify G with roots of A(T)=T^{p^s}+T in the finite field.
domain assumption K is algebraically closed in F (full constant field).
Standing hypothesis for the descent problem; ensures the descent problem is stated correctly and the constant field of subextensions remains K.
domain assumption The additive polynomial A(T)=T^{p^s}+T has coefficients in the base field k, so the Frobenius action on G is well defined.
Section 3.2 explicitly assumes A(x) in k[x]; Section 4 takes A(T)=T^{p^s}+T, which lies in F_p[T] subset k[T], so the action Theta(sigma) is defined.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the Weil descent of Artin-Schreier algebraic function fields over finite fields." pith.science (2026). https://pith.science/paper/4AWI57DO

@misc{pith2026250521656,
  author       = {Pith},
  title        = {Pith review of: On the Weil descent of Artin-Schreier algebraic function fields over finite fields},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4AWI57DO}},
  note         = {Machine review of arXiv:2505.21656}
}

abstract

Let us consider a generalized Artin-Schreier algebraic function field extension $F$ of the rational function field $\F_{p^n}(x)$ defined over the finite field extension $K=\F_{p^n}$ of the prime field $\F_p$. We assume that $K$ is algebraically closed in $F$. We give general results on the descent over the fields $k= \F_{p^t}$ for $t$ dividing $n$. Then, we completely handle the bi-cyclic case of the descent over the fields $k_1=\F_{p}$ and $k_2= \F_{p^2}$ of all the sub-extensions of $F$ defined over $\F_{p^4}$. We give explicit examples with small prime numbers $p$.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

Ballet, Low Increasing Tower of Algebraic Function Fields and Bilin ear Complexity of Multiplication in Any Extension of Fq

S. Ballet, Low Increasing Tower of Algebraic Function Fields and Bilin ear Complexity of Multiplication in Any Extension of Fq. Finite Fields and Their Applications, 9 (2003), 472- 478

work page 2003
[2]

Ballet, D

S. Ballet, D. Le Brigand, R. Rolland, On an application of the deﬁnition ﬁeld descent of a tower of function ﬁelds. Proceedings of the Conference Arithmeti c, Geometry and Coding Theory (AGCT 2005), Soci´ et´ e Math´ ematique de France, s´ er. S´ eminaires et Congr` es, 21, p. 187–203, 2009

work page 2005
[3]

Ballet, R

S. Ballet, R. Rolland, Multiplication Algorithm in a Finite Field and Tensor Rank o f the Multiplication. Journal of Algebra, 272 (2004), 173-185

work page 2004
[4]

Deolalikar, Determining irreducibility and ramiﬁcation groups for an a dditive extension of the rational function ﬁeld

V. Deolalikar, Determining irreducibility and ramiﬁcation groups for an a dditive extension of the rational function ﬁeld. Journal of Number Theory, 97 (20 02), no. 2, 269-286

work page
[5]

Garcia, H

A. Garcia, H. Stichtenoth, Elementary Abelian p-Extensions of Algebraic Function Fields. Manuscripta Mathematica, 72 (1991), 67-79

work page 1991
[6]

Garcia, H

A. Garcia, H. Stichtenoth, A tower of Artin-Schreier extensions of function ﬁelds atta ining the Drinfeld-Vladut bound. Inventiones Mathematicae, 121 (1995), 211-222

work page 1995
[7]

Garcia, H

A. Garcia, H. Stichtenoth, A class of polynomials over ﬁnite ﬁelds. Finite Fields and Th eir Applications, 5 (1999), no. 4, 424-435

work page 1999
[8]

Goss, Basic structures of function ﬁelds arithmetic

D. Goss, Basic structures of function ﬁelds arithmetic. Springer 19 91

work page

Show all 14 references

[9]

R. Lidl, H. Niederreiter, Introduction to ﬁnite ﬁelds and their applications. Cambri dge Univ. Press, 1994 (revised version)

1994
[10]

Sadi, Descente Eﬀective du Corps de D´ eﬁnition des Revˆ etements

B. Sadi, Descente Eﬀective du Corps de D´ eﬁnition des Revˆ etements. Th` ese pr´ esent´ ee ` a l’Universit´ e de Lille 1, 7 mai 1999

1999
[11]

Serre, Groupes Alg´ ebriques et Corps de Classes

J.-P. Serre, Groupes Alg´ ebriques et Corps de Classes. Hermann, Paris, 1 959
[12]

Shparlinski, M.A

I.E. Shparlinski, M.A. Tsfasman, S.G. Vladut, Curves with Many Points and Multiplication in Finite Fields. Lectures Notes in Mathematics, Springer- Verlag, 1518 (1992), 145-169

1992
[13]

Stichtenoth, Algebraic Function Fields and Codes

H. Stichtenoth, Algebraic Function Fields and Codes. Springer Universitex t, Berlin/Heidelberg/New York, 1993

1993
[14]

Weil, The Field of Deﬁnition of a Variety

A. Weil, The Field of Deﬁnition of a Variety. American Journal Of Math ematics, 78 (1956), 509-524. Aix Marseille Univ, CNRS, Centrale Marseille, I2M, Marseille, France, Institut de Math´ematiques de Marseille, Case 907, 13288 Marseille cedex 9. Email address : stephane.ballet...

1956

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Ballet, Low Increasing Tower of Algebraic Function Fields and Bilin ear Complexity of Multiplication in Any Extension of Fq

S. Ballet, Low Increasing Tower of Algebraic Function Fields and Bilin ear Complexity of Multiplication in Any Extension of Fq. Finite Fields and Their Applications, 9 (2003), 472- 478

work page 2003

[2] [2]

Ballet, D

S. Ballet, D. Le Brigand, R. Rolland, On an application of the deﬁnition ﬁeld descent of a tower of function ﬁelds. Proceedings of the Conference Arithmeti c, Geometry and Coding Theory (AGCT 2005), Soci´ et´ e Math´ ematique de France, s´ er. S´ eminaires et Congr` es, 21, p. 187–203, 2009

work page 2005

[3] [3]

Ballet, R

S. Ballet, R. Rolland, Multiplication Algorithm in a Finite Field and Tensor Rank o f the Multiplication. Journal of Algebra, 272 (2004), 173-185

work page 2004

[4] [4]

Deolalikar, Determining irreducibility and ramiﬁcation groups for an a dditive extension of the rational function ﬁeld

V. Deolalikar, Determining irreducibility and ramiﬁcation groups for an a dditive extension of the rational function ﬁeld. Journal of Number Theory, 97 (20 02), no. 2, 269-286

work page

[5] [5]

Garcia, H

A. Garcia, H. Stichtenoth, Elementary Abelian p-Extensions of Algebraic Function Fields. Manuscripta Mathematica, 72 (1991), 67-79

work page 1991

[6] [6]

Garcia, H

A. Garcia, H. Stichtenoth, A tower of Artin-Schreier extensions of function ﬁelds atta ining the Drinfeld-Vladut bound. Inventiones Mathematicae, 121 (1995), 211-222

work page 1995

[7] [7]

Garcia, H

A. Garcia, H. Stichtenoth, A class of polynomials over ﬁnite ﬁelds. Finite Fields and Th eir Applications, 5 (1999), no. 4, 424-435

work page 1999

[8] [8]

Goss, Basic structures of function ﬁelds arithmetic

D. Goss, Basic structures of function ﬁelds arithmetic. Springer 19 91

work page