Pith. sign in

REVIEW 1 major objections 5 minor 4 references

On the Distribution of Class Groups of Abelian Extensions

T0 review · 1 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper proves that, over function fields, the weighted moments of the good part of the p-class group in abelian extensions equal $1/|M|$, matching the random finite module model.

desk verdict Substantial new results on class group distributions for abelian extensions, with a narrow but load-bearing gap in Lemma 10.2 that needs verification before the weighted moment theorem is accepted. read the letter →

arxiv 2411.19318 v1 pith:RLXFPKZH submitted 2024-11-28 math.NT

Yuan Liu This is my paper · ORCID
classification math.NT MSC 11R2911R3211R4511R5814H30
keywords classgroupsabelianextensionsp-primarygroupgenustheoryfunctionfieldstame-covermodulispacesmomentconjecturesramificationtypes
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

For a finite abelian group $\Gamma$, the paper asks how the $p$-primary part of the class group of a $\Gamma$-extension of $\mathbb{Q}$ or $\mathbb{F}_q(t)$ behaves as the extension varies. Its answer is a split: for each primitive idempotent $e$ of $\mathbb{Q}_p[\Gamma]$, the module $e\mathrm{Cl}(K)=e\mathbb{Z}_p[\Gamma]\otimes_{\mathbb{Z}_p[\Gamma]}\mathrm{Cl}(K)[p^\infty]$ contains a forced 'genus' quotient $e\mathrm{Cl}(K)/(I_e\cdot e\mathrm{Cl}(K))$ whose size grows with the number of suitably ramified primes, and a complementary 'good' module $I_e\cdot e\mathrm{Cl}(K)$ that should be random. The paper proves the moment side of this randomness in the function-field case: with a weight function coming from the bad quotient, the average of $\#\mathrm{Sur}_\Gamma(I_e\cdot e\mathrm{Cl}(K),M)$ over totally real $\Gamma$-extensions of $\mathbb{F}_q(t)$ with $\mathrm{rDisc}(K)=q^n$, for $0\le n\le N$, tends to $1/|M|$ in the iterated limit $q\to\infty$ then $N\to\infty$. This recovers the classical class-group heuristic in the tame case and the cyclic $p$-extension distribution in the case $\Gamma=\mathbb{Z}/p\mathbb{Z}$, and it identifies the ideal $I_e$ explicitly.

What carries the argument

The load-bearing object is the discrete valuation ring $e\mathbb{Z}_p[\Gamma]$ attached to a primitive idempotent $e$ of $\mathbb{Q}_p[\Gamma]$, together with the ideal $I_e$ defined as the intersection, over nontrivial $\gamma\in\Gamma$, of the images in $e\mathbb{Z}_p[\Gamma]$ of the ideals $(1-\gamma,\,1+\gamma+\cdots+\gamma^{|\gamma|-1})$. This ring classifies finite $e\mathbb{Z}_p[\Gamma]$-modules by their $I$-ranks and separates each $e\mathrm{Cl}(K)$ into the bad quotient $e\mathrm{Cl}(K)/(I_e\cdot e\mathrm{Cl}(K))$ and the good module $I_e\cdot e\mathrm{Cl}(K)$. The weighted-moment proof counts points on tame Galois-cover moduli spaces: extensions with a prescribed module $H$ correspond to points of these moduli spaces, an identity in Proposition 9.3 converts surjections onto $H$ into surjections onto $I_eH$ times a weight factor, and a comparison lemma for point counts identifies the dominant contribution as $1/|M|$.

What would settle it

Compute the left-hand side of (1.3) for a small explicit case, such as $\Gamma=\mathbb{Z}/2\mathbb{Z}$, $p=2$, and a finite module $M$, for $q\equiv 3\bmod 4$ and $q\equiv 1\bmod 4$; Theorem 1.3 predicts respective limits $1/|M|$ and $|(\wedge^2M)[2^{v-1}]|/|M|$, so a single deviation for one module or base field would falsify the moment claim.

Watch

Extended reading notes

Core claim

The central claim is that the $p$-primary class group of an abelian $\Gamma$-extension, as a $\mathbb{Z}_p[\Gamma]$-module, decomposes into a statistically infinite part forced by ramification and a statistically finite part obeying the predicted moments. Concretely, each primitive idempotent $e$ of $\mathbb{Q}_p[\Gamma]$ gives a discrete valuation ring $e\mathbb{Z}_p[\Gamma]$, and there is a canonical ideal $I_e$ such that the quotient $e\mathrm{Cl}(K)/(I_e\cdot e\mathrm{Cl}(K))$ has $I_e$-rank bounded below by the number of primes of certain ramification types, so its average is infinite. The complementary module $I_e\cdot e\mathrm{Cl}(K)$ is conjectured to be equidistributed by the unique probability measure with $M$-moment $1/|M|$; Theorem 1.2(2) proves the weighted version of this moment statement over function fields for every nontrivial $e$. The weight in the average is exactly the homomorphism count from the bad quotient to a fixed module, so the result says that the bad part, despite being infinite on average, does not change the moments of the good part.

Load-bearing premise

The weighted-moment theorem rests on the claim that a technical admissibility condition in the cited moduli-space count can be dropped; if that condition is genuinely needed, the point counts can differ and the moment equality could fail.

Editorial extensions

If this is right

  • For every nontrivial primitive idempotent $e$, the weighted moment of the good part equals the moment of the unique random finite module measure, so the good part is fully determined by these moments.
  • The bad part $e\mathrm{Cl}(K)/(I_e\cdot e\mathrm{Cl}(K))$ has infinite average rank, so any unweighted moment statement must fail; the weighting in (1.3) is essential in the function-field count.
  • When $p\nmid |\Gamma|$, the weight function is constant and the theorem recovers the moment version of the standard class-group heuristic; when $\Gamma=\mathbb{Z}/p\mathbb{Z}$, it gives the predicted distribution of $(1-\gamma)\mathrm{Cl}(K)[p^\infty]$.
  • The trivial idempotent component is bounded by $|\wedge^2\Gamma_p|$, so the full class group's growth is carried by the nontrivial $e$-components and by the kernel of the map to $\bigoplus_e e\mathrm{Cl}(K)$, whose average rank is infinite when $p^2\mid |\Gamma|$.

Reading between the lines

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

  • The same weighted-moment strategy is pointed toward other bad-prime settings, such as base fields containing the relevant roots of unity, where moduli-space moments diverge; carrying it out would require tracking central-extension corrections of the kind isolated in the comparison lemma.
  • The bound $|I_{e_0}(e_0\mathrm{Cl}(K))|\le |\wedge^2\Gamma_p|$ suggests a general pattern: in wild abelian extensions, the trivial-representation component is controlled by the wedge square of the maximal $p$-power subgroup of $\Gamma$, so it vanishes when that subgroup is cyclic.
  • A direct numerical check over $\mathbb{F}_q(t)$ for small $q$ and $N$ could test whether the weight function and the good part are asymptotically independent; if not, the unweighted conjecture would fail in a detectable way.
  • The infinite average kernel when $p^2\mid |\Gamma|$ indicates that the idempotent decomposition misses a statistically large piece exactly in the wild case, so full distribution results may need to parametrize this kernel as well.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 5 minor

Summary. The paper studies the distribution of the p-part of class groups Cl(K) as K varies over totally real Galois extensions of Q or F_q(t) with a fixed finite abelian Galois group Γ. For each primitive idempotent e of Q_p[Γ] the author constructs a discrete valuation ring eZ_p[Γ] and an ideal I_e, and proves a lower bound (an analogue of genus theory) for the rank of the bad part eCl(K)/(I_e eCl(K)). The main results are: Theorem 1.2(1), that the average I_e-rank is infinite; Theorem 1.2(2), a weighted moment statement in the function field case asserting that the weighted average of #Sur_Γ(I_e eCl(K), M) is 1/|M|; Theorem 1.3, a refinement for Γ=Z/2Z and p=2; and Theorem 1.4 on the kernel of Cl(K) → ⊕_e eCl(K). The paper also states Conjecture 12.2 giving the unweighted moment and probability distribution for I_e eCl(K), and shows that the conjectures specialize to the Cohen-Lenstra-Martinet and Gerth heuristics when p ∤ |Γ| and when Γ=Z/pZ respectively. The proofs combine cohomological methods for presentations of Galois groups with restricted ramification and Hurwitz-space point counting in the function field setting.

Significance. If the results are correct, this is a substantial generalization of the Cohen-Lenstra-Gerth heuristic framework to all finite abelian Galois groups, including the difficult bad prime cases where p divides |Γ|. The weighted-moment technique introduced here is designed to handle infinite moments coming from Hurwitz-space counts and is likely to be of independent use. The paper gives detailed proofs of the main theorems, and the agreement of the conjectures with the known Cohen-Lenstra-Martinet and Gerth predictions provides strong internal consistency. The unweighted Conjecture 12.2 is honestly labeled as a conjecture, which is appropriate since the proof is not attempted. The main concern is a load-bearing point in the proof of the weighted moment theorem (Theorem 1.2(2)) where a hypothesis from an external lemma is removed without proof; this is discussed in the major comments.

major comments (1)
  1. [§10.2, proof of Lemma 10.2(2), footnote 3] The proof of Lemma 10.2(2) invokes [LWZB24, Lemma 12.10] to obtain Schur coverings S_i → G_i and S_{Γ'} → Γ' with the property that the kernel of f|ker(S_i→G_i) has p-power order; this property is then used in (10.10) to equate nr_{q-1} values and conclude b(G1,c1,q,n) = b(G2,c2,q,n). For the groups G_i = (H_i × Γ_p) ⋊ Γ' that arise in Proposition 10.3, the admissibility hypothesis of [LWZB24, Lemma 12.10] is not satisfied when Γ_p is nontrivial. The footnote asserts that the admissibility condition can be removed because it is not used in the proof, but no argument or reference is supplied. This is load-bearing: if the p-power-order property fails for these non-admissible groups, the equality of b-values in Lemma 10.2(2) is unsupported, and with it Proposition 10.3 and Theorem 1.2(2) collapse. The author should either give a complete proof of the removal, or modify the argument to avoid relying on the admissibility hypothesis in the non-admissible cases.
minor comments (5)
  1. [§4.4, Lemma 4.4(3)] The proof of Lemma 4.4 iteratively adds primes p not in S ∪ T'(Q) to make B vanish, but it does not ensure that the added primes avoid ∪_{ℓ|p|Γ|} S_ℓ(Q). Condition (3) counts only primes outside that union. The argument appears to establish only the upper bound #(S minus (∪_{ℓ|p|Γ|} S_ℓ(Q) ∪ S ∪ T)) ≤ dim B / dim End, which is all that is later used in §6.3 and §7, but the stated equality is not justified. The lemma statement should either be weakened to an upper bound or the proof should be completed.
  2. [§10.2, proof of Lemma 10.2(2), Claim 2] The assertion 'Since (Hi)Γ is an abelian p-group and p∤q−1, we have m>1' is not literally correct for elements g whose image γ ∈ Γ′ has order dividing q−1: for such γ one has γ^q = γ and m = 1. The conclusion of Claim 2 still holds in that case because (1−γ) is a unit on eZ_p[Γ] so the H-coinvariant part is trivial, but the proof should be rephrased to treat the cases γ ∈ Γ_p and γ ∈ Γ′ separately.
  3. [§2, Lemma 2.6] In the proof of Lemma 2.6, the phrase 'for e1 ≠ e2 in Idem(Fp)' should read 'in Idem(A)' to match the notation of the lemma.
  4. [§1.4 and Theorem 1.2(2)] The iterated-limit notation in §1.4 defines lim_{x→∞} lim_{y→∞} for a two-variable function, but Theorem 1.2(2) and Proposition 10.3 impose additional restrictions on q (p∤q(q−1), gcd(q,|Γ|)=1). It would be helpful to state explicitly that the inner limit is taken over q satisfying these restrictions.
  5. [§2, Definition 2.9] In Definition 2.9, 'the maximal interger r' contains a typo; it should be 'integer'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the weighted-moment theorem rests on published Hurwitz-space counting and explicit algebraic identities, not on its own conclusion.

full rationale

Walking the derivation chain of Theorem 1.2(2), the equality in (1.3) is obtained by combining Proposition 9.3, an algebraic identity of the form #Sur_eZp[Γ](M,H)=w(M,H)#Sur_eZp[Γ](IM,IH), with Proposition 10.3, which reduces the ratio of #Sur_eZp[Γ](Cl(K),H1) to #Sur_eZp[Γ](Cl(K),H2) to equality of Hurwitz-space main terms b(G1,c1,q,n)=b(G2,c2,q,n), and with [LWZB24, Proposition 12.7 and Corollary 12.9] for point-count asymptotics. None of these inputs is equivalent to (1.3): the theorem's content is the asymptotic comparison of two Hurwitz counts, and Proposition 9.3 only rewrites one count as the weighted moment of another. The weight function is not a fitted parameter; it is defined by the choice H2=(eZp[Γ]/I_e)^r, and the theorem proves the count comparison for H1 versus H2. Conjecture 12.2 is explicitly labeled a conjecture and is not used to prove Theorem 1.2(2); its agreement with Cohen-Lenstra-Martinet and Gerth is an external consistency check, not a circular reuse. The load-bearing step that deserves scrutiny is Lemma 10.2(2), where the proof removes the admissibility hypothesis from [LWZB24, Lemma 12.10] with the assertion 'this condition can be removed because it is not used in the proof.' If that assertion were false, the equality of b-values and hence Theorem 1.2(2) could fail; this is a correctness and verification risk, not a circularity, because the cited lemma is independent published work and the paper never assumes the conclusion it is trying to prove. I find no step in which a 'prediction' is equivalent to its input by construction or by self-citation.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The paper's central claims rest on standard class field theory, on quoted counting theorems (Wood, Sawin-Wood, Hurwitz space counts), and on one ad-hoc unproved relaxation of a hypothesis in [LWZB24, Lemma 12.10]. No numerical constants are fitted to data; the weight function is a deliberate construction, not a fitted parameter.

assumptions (6)
  • standard math Global class field theory identifies Cl(K) with the Galois group of the maximal abelian unramified (and in function fields, split-at-∞) extension of K.
    Used throughout, e.g. in Section 3 definitions of Cl(K) and in Proposition 10.4, to translate class group surjections into unramified extensions.
  • standard math Wood's theorem [Woo10] on the distribution of local behaviors in abelian extensions, used to count Γ-extensions with prescribed local specifications.
    Appendix A uses [Woo10, Theorem 2.1] to prove Theorem A.1, which feeds Theorem 3.8 and Corollary 3.6.
  • standard math Sawin-Wood [SW22] moment determinacy theorem for random objects in a category.
    Proposition 12.1 uses [SW22, Theorem 1.2 and Lemma 6.3] to construct and identify the unique Cohen-Lenstra type measure.
  • standard math Hurwitz space point counts and braid monodromy main terms from [LWZB24, Proposition 12.7 and Lemma 12.10].
    Sections 10 and 11 rely on these to estimate #Hur^n_{G,c}(F_q); the same-author result is published, but the paper modifies Lemma 12.10 by dropping admissibility.
  • ad hoc to paper The admissible Γ-group hypothesis in [LWZB24, Lemma 12.10] can be dropped for groups G = H ⋊ Γ where H is a finite eZ_p[Γ]-module.
    Assumed in the proof of Lemma 10.2 without a verification in this paper; it is load-bearing for Lemma 10.2, Proposition 10.3, and Theorem 1.2(2).
  • domain assumption The function field moment sums are nonzero and the iterated limit exists as defined in Section 1.4.
    Theorem 1.2(2) is stated only for the iterated q→∞, N→∞ limit with p∤q(q-1) and gcd(q,|Γ|)=1; the result does not address fixed q or number field moments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Distribution of Class Groups of Abelian Extensions." pith.science (2026). https://pith.science/paper/RLXFPKZH

@misc{pith2026241119318,
  author       = {Pith},
  title        = {Pith review of: On the Distribution of Class Groups of Abelian Extensions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RLXFPKZH}},
  note         = {Machine review of arXiv:2411.19318}
}
abstract

Given a finite abelian group $\Gamma$, we study the distribution of the $p$-part of the class group $\operatorname{Cl}(K)$ as $K$ varies over Galois extensions of $\mathbb{Q}$ or $\mathbb{F}_q(t)$ with Galois group isomorphic to $\Gamma$. We first construct a discrete valuation ring $e\mathbb{Z}_p[\Gamma]$ for each primitive idempotent $e$ of $\mathbb{Q}_p[\Gamma]$, such that 1) $e\mathbb{Z}_p[\Gamma]$ is a lattice of the irreducible $\mathbb{Q}_p[\Gamma]$-module $e\mathbb{Q}_p[\Gamma]$, and 2) $e\mathbb{Z}_p[\Gamma]$ is naturally a quotient of $\mathbb{Z}_p[\Gamma]$. For every $e$, we study the distribution of $e\operatorname{Cl}(K):=e\mathbb{Z}_p[\Gamma] \otimes_{\mathbb{Z}_p[\Gamma]} \operatorname{Cl}(K)[p^{\infty}]$, and prove that there is an ideal $I_e$ of $e\mathbb{Z}_p[\Gamma]$ such that $e\operatorname{Cl}(K) \otimes (e\mathbb{Z}_p[\Gamma]/I_e)$ is too large to have finite moments, while $I_e \cdot e\operatorname{Cl}(K)$ should be equidistributed with respect to a Cohen--Lenstra type of probability measure. We give conjectures for the probability and moment of the distribution of $I_e\cdot e\operatorname{Cl}(k)$, and prove a weighted version of the moment conjecture in the function field case. Our weighted-moment technique is designed to deal with the situation when the function field moment, obtained by counting points of Hurwitz spaces, is infinite; and we expect that this technique can also be applied to study other bad prime cases. Our conjecture agrees with the Cohen--Lenstra--Martinet conjecture when $p\nmid |\Gamma|$, and agrees with the Gerth conjecture when $\Gamma=\mathbb{Z}/p\mathbb{Z}$. We also study the kernel of $\operatorname{Cl}(K) \to \bigoplus_e e\operatorname{Cl}(K)$, and show that the average size of this kernel is infinite when $p^2\mid |\Gamma|$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references · 2 canonical work pages

  1. [13]

    Wright, Distribution of discriminants of abelian extensions , Proc

    MR4240808 [Wri89] David J. Wright, Distribution of discriminants of abelian extensions , Proc. London Math. Soc. (3) 58 (1989), no. 1, 17–50. MR969545 [WW21] Weitong Wang and Melanie Matchett Wood, Moments and interpretations of the Cohen-Lenstra-Martine t heuristics, Comment. Math. Helv. 96 (2021), no. 2, 339–387. MR4277275 Department of Mathematics, Uni...

    work page 1989

  2. [1977]

    preprint, arXiv:2207.05674

    MR450380 [Smi22] Alexander Smith, The distribution of ℓ∞-Selmer groups in degree ℓ twist families (2022). preprint, arXiv:2207.05674. [Sta18] The Stacks Project Authors, Stacks Project,

    arXiv 2022

  3. [1998]

    Presentations of Galois groups of maximal extensions with restricted ramification

    Basi c representation theory of finite groups and associative algebras. MR1644252 [BW17] Nigel Boston and Melanie Matchett Wood, Non-abelian Cohen-Lenstra heuristics over function fields , Compos. Math. 153 (2017), no. 7, 1372–1390. MR3705261 [CL84] H. Cohen and H. W. Lenstra Jr., Heuristics on class groups of number fields , Number theory, Noordwijk- erhout...

    work page Pith review arXiv 2017

  4. [2018]

    preprint, arXiv:2210.06279

    [SW22] Will Sawin and Melanie Matchett Wood, The moment problem for random objects in a category (2022). preprint, arXiv:2210.06279. [Woo10] Melanie Matchett Wood, On the probabilities of local behaviors in abelian field exte nsions, Compos. Math. 146 (2010), no. 1, 102–128. MR2581243 [Woo21] , An algebraic lifting invariant of Ellenberg, Venkatesh, an d W...

    arXiv 2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.