Pith. sign in

REVIEW 3 major objections 4 minor 2 references

A Dieudonn\'e theory for analytic p-divisible groups and applications to Shimura varieties

T0 review · 3 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Families of analytic p-divisible groups are classified by a single linear map f_G — this paper proves the equivalence.

desk verdict A genuinely strong family-level Dieudonné theory paper whose central theorem leans on unverified companion-paper lemmas; referee with [Ger26] in hand. read the letter →

arxiv 2602.05764 v2 pith:2QTM6V56 submitted 2026-02-05 math.AG math.NT

Lucas Gerth This is my paper · ORCID
classification math.AGmath.NT MSC 14L0514G2211G18
keywords analyticp-divisiblegroupsHodge–Tatetriplesp-adicHodgetheoryDieudonnélocalshtukasShimuravarietiesCartierduality
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

Over any good adic space S over Q_p — the class that includes perfectoid and seminormal rigid spaces — the paper proves that a family of analytic p-divisible groups is completely encoded by linear algebra: a Z_p-local system L, a vector bundle E, and one morphism f between their pullbacks to the v-site. In concrete terms, L is the Tate module of the group, E is its Lie algebra, and f is the derivative of a natural pairing; the group can be rebuilt from this triple, and its logarithm exact sequence is a pullback of a standard sequence. From this equivalence the paper extracts a Dieudonné theory: dualizable groups match minuscule local shtukas and vector bundles on the relative Fargues–Fontaine curve, with Cartier duality built in. It then uses this to reinterpret local Shimura varieties of EL/PEL type as moduli spaces of such groups with extra structure, and to rewrite the Hodge–Tate period map on global PEL Shimura varieties as the passage from an abelian variety to its topologically p-torsion subgroup. The proof imports a technical representability lemma from the author's companion paper, and the main theorem stands or falls with it.

What carries the argument

The load-bearing object is the Hodge–Tate triple (L, E, f), and within it the single map f_G: Lie(G)⊗O_{S_v} → T_pG(-1)⊗O_{S_v}. The map is constructed as the derivative at the identity of the analytic Weil pairing e: G × T_pG^∨(1) → G_m⟨p^∞⟩, which uniquely extends the classical pairing on p-power torsion. The proof that this linear datum determines the whole group runs through the logarithm exact sequence 0 → G[p^∞] → G → Lie(G) → 0 and a representability lemma imported from the companion paper, which guarantees that certain pullbacks of v-sheaves are analytic p-divisible groups. All later constructions — Cartier duality, minuscule local shtukas, the vector bundle E(G), and the moduli appl

What would settle it

Take a good adic space S that is not perfectoid, such as a seminormal rigid space, pick a tuple (L, E, f) with f = 0, and check whether the v-sheaf pullback defined in Theorem 3.13 is representable by an analytic p-divisible group; a negative answer falsifies the companion lemma and with it the main equivalence. Alternatively, find two non-isomorphic analytic p-divisible groups over some S with the same Tate module, the same Lie algebra, and the same map f_G — the theorem asserts they cannot exist.

Watch

Extended reading notes

Core claim

The central claim is Theorem 3.13: for every good adic space S over Q_p there is a canonical equivalence between the category of analytic p-divisible groups G → S and the category of triples (L, E, f) in which L is a Z_p-local system on the v-site of S, E is a vector bundle on the étale site of S, and f: E⊗O_{S_v} → L(-1)⊗O_{S_v} is a morphism of v-vector bundles. If G corresponds to (L, E, f), then L = T_pG, E = Lie(G), and f = f_G is the Hodge–Tate map obtained as the derivative of the analytic Weil pairing; moreover the logarithm sequence of G is the pullback of the product sequence for T_pG(-1)⊗G_m⟨p^∞⟩. The paper shows this equivalence is compatible with base change and uses it as the f

Load-bearing premise

The main theorem relies on a technical lemma from the author's companion paper guaranteeing that certain pullbacks of v-sheaves are representable by analytic p-divisible groups; that lemma is not proved in this preprint, and if it fails for any good adic space the whole edifice falls.

Editorial extensions

If this is right

  • Analytic p-divisible groups over good adic bases are effectively linear-algebraic: isogenies, extensions, and Cartier duality are all encoded in the single map f_G, so computations can be done with local systems and vector bundles.
  • Dualizable analytic p-divisible groups are equivalent to minuscule local shtukas, hence to vector bundles on the relative Fargues–Fontaine curve with minuscule modifications; this gives a Dieudonné theory with a built-in Cartier dual.
  • For a proper smooth family, the topologically p-torsion part of the relative Picard variety is an analytic p-divisible group up to a maximal open subgroup; for abeloid varieties its Cartier dual is the topologically p-torsion subgroup of the dual abeloid variety.
  • Local Shimura varieties of EL and PEL type are moduli spaces of dualizable analytic p-divisible groups with extra structure, valid for arbitrary level subgroups and independent of integral models.
  • The Hodge–Tate period map on PEL Shimura varieties over Q sends an abelian variety to its topologically p-torsion subgroup, and the first de Rham–Fargues–Fontaine cohomology of a proper smooth rigid space is recovered as E(H)(1) for the maximal analytic p-divisible subgroup of its topologically p-torsion Picard variety.

Reading between the lines

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

  • The reduction to (L, E, f) suggests the whole category of analytic p-divisible groups over a good adic space is controlled by an étale vector bundle and a v-local system; a natural test is whether every extension of such vector bundles by such local systems arises from an analytic p-divisible group, and whether the equivalence extends beyond good adic spaces where the diamond functor is no longer
  • The functor E(−) on the relative curve depends only on the p-adic universal cover, not on the group itself; since non-isogenous abeloid varieties can have isomorphic universal covers, the framework yields identifications between their first de Rham homology groups — a plausible route to de Rham period domains outside the good-reduction locus.
  • The companion-paper representability results are the linchpin; if they extend to perfectoid bases over Spd(Z_p), the same equivalence should produce a minuscule variant of the local-shtuka stack in mixed characteristic, extending the reinterpretation of the Hodge–Tate period map beyond characteristic zero.
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.

Referee Report

3 major / 4 minor

Summary. The paper develops a relative analytic Dieudonné theory for families of analytic p-divisible groups over good adic spaces over Q_p. The central theorem (Thm. 1.1 / Thm. 3.13) asserts a canonical equivalence between analytic p-divisible groups G→S and triples (L,E,f), where L is a Z_p-local system on S_v, E is a vector bundle on S_ét, and f:E⊗O_{S_v}→L(−1)⊗O_{S_v}. The paper then constructs a sheaf E(G) on the relative Fargues–Fontaine curve, compares it with minuscule shtukas, defines analytic Dieudonné crystals over good adic spaces, and derives applications to Hodge–Tate period maps, local Shimura varieties of EL/PEL type, and higher direct images. A substantial part of the argument is explicitly imported from the companion paper [Ger26], including the representability of the fiber product defining the inverse equivalence and the diamantine higher direct image results.

Significance. If the imported companion results are correct, this is a substantial contribution: it gives a complete linear-algebraic description of analytic p-divisible groups in families, recovers Fargues' theorem at a point and Scholze–Weinstein over Spa(C,O_C), and is benchmarked against Anschütz–Le Bras prismatic Dieudonné theory and Heuer's work. The paper contains no fitted parameters, and the external benchmarks provide meaningful consistency checks. The main risk is not internal inconsistency but the heavy reliance on [Ger26] for load-bearing lemmas that are not reproduced or proved here.

major comments (3)
  1. [Theorem 3.13 (§3.3)] The proof of the central equivalence depends twice on [Ger26, Lemma 3.38]: first to assert that the fiber product F(L,E,f) is representable by an analytic p-divisible group, and again to assert that the short exact sequence defining F(L,E,f) agrees with the logarithm sequence. These are precisely the two properties that make the inverse functor well-defined and the unit/counit of the equivalence valid. The lemma is not stated or proved in this manuscript. If it has hidden hypotheses (for example perfectoid base or a local direct-summand condition on f), then Theorem 3.13, and with it Theorems 5.23, 5.26, 6.7, and 7.8, would fail. This is a gap in support rather than an observed contradiction, but it is load-bearing and must be addressed, either by proving [Ger26, Lemma 3.38] in this paper or by reproducing its full statement and proof.
  2. [Theorem 4.10 (§4.2)] Theorem 4.10 is quoted verbatim from [Ger26, Prop. 3.16 and Prop. 3.39] and supplies the diamantine higher direct image statements used in Theorem 4.11, Corollary 4.13, and Proposition 6.4. In particular, the representability of R^nπ_{ét,*}G[p^m], the exact sequence (4.7), and the isomorphism (4.9) are all imported. These are not peripheral facts: without them, the Picard-variety applications (Cor. 4.13), the abeloid duality theorem (Cor. 4.20), and the v-descent argument in Prop. 6.4 do not go through. The manuscript should either include the statements and proofs of these results or state clearly that the paper is conditional on them.
  3. [Theorem 5.26 and Definition 5.25 (§5.3)] The descent from perfectoid bases to arbitrary good adic spaces is asserted in a single sentence: after reducing to a functorial rule on perfectoid test objects whose Lie algebras and dual Lie algebras come from étale vector bundles, the proof says the equivalence 'follows directly from Theorem 5.23'. Given Remark 3.17 explicitly notes that analytic p-divisible groups do not satisfy v-descent over arbitrary good adic spaces, this is not automatic. The proof should spell out how the analytic Dieudonné crystal condition supplies the missing descent data, construct the inverse functor on an arbitrary good S, and verify the unit/counit. As written, this is a load-bearing step in the paper's main Dieudonné-theoretic claim.
minor comments (4)
  1. [§3.3, proof of Theorem 3.13] The final paragraph reads 'Proposition 3.16 shows that F^{-1}∘F=id. To conclude that F^{-1}∘F=id...' The second expression should presumably be F∘F^{-1}=id; as written it is a typo that makes the argument confusing.
  2. [§2.3, Lemma 2.15] The proof of Lemma 2.15 refers to 'Lemma 2.9(3)', but Lemma 2.9 as stated has only items (1) and (2). The reference should be corrected, and if a third item is intended it should be stated.
  3. [Example 3.32] The example asserts the existence of a formal elliptic curve E→Spf(C^+) with E_{k(s)} supersingular and E_{k(u)} ordinary for a higher-rank valuation subring C^+, but no construction or reference is supplied. Since this example is used to show that the adic generic fiber functor is not fully faithful over such bases, please add a reference or a proof.
  4. [Throughout] There are several typos: 'independant' in the paragraph after Theorem 1.2, 'anyltlic' in §1.1 ('dualizable anylticp-divisible group'), and 'Rapoport–Zink envisioned' in the introduction. These should be corrected.

Circularity Check

2 steps flagged · score 4.0 of 10

Theorem 3.13 is concluded by invoking the author's companion-paper lemma [Ger26, Lemma 3.38]; the central equivalence is load-bearing self-citation, though not a definitional reduction.

  1. self citation load bearing [§3.3, proof of Theorem 3.13]
    "Hence it follows from [Ger26, Lemma 3.38] that G is representable by an analytic p-divisible group, so that the functor F is well-defined. ... This is [Ger26, Lemma 3.38], which concludes."

    The proof of the paper's foundational equivalence is not completed inside the paper: the inverse functor's well-definedness and the identity with the logarithm sequence are both asserted to be exactly [Ger26, Lemma 3.38]. Thus, as written, Theorem 3.13 reduces to that lemma rather than to an argument contained here. Since [Ger26] is the author's companion paper and is not reproduced, the central claim is supported by a load-bearing self-citation rather than by a first-principles derivation.

  2. self citation load bearing [§4.2, Theorem 4.10]
    "We will use the following results from [Ger26]. Theorem 4.10. ([Ger26, Prop. 3.16, Prop. 3.39]) Let π:X→S be a proper smooth morphism of seminormal rigid spaces over K and let G be an analytic p-divisible X-group. Assume either that S=Spa(K,K^+) or that the Lie algebra g is the pullback of a vector bundle on S."

    The higher-direct-image representability and exact-sequence results are imported verbatim from the same companion paper and are then used for Corollary 4.13 and Proposition 6.4. This is a second load-bearing self-citation: if those [Ger26] results fail or require extra hypotheses, the Picard-variety applications and the v-descent argument in Proposition 6.4 do not go through. It is not a definitional reduction, but it makes later results depend on the same un-reproduced self-citation chain.

full rationale

This is pure mathematics with no fitted parameters, so the fitted-input-as-prediction pattern does not apply. The central content is not defined in terms of its conclusion, and there is no visible renaming of a known empirical pattern. The main issue is that Theorem 3.13, from which nearly all later results are derived, is concluded by citing [Ger26, Lemma 3.38] for both the representability of the constructed fiber product and the identification of its logarithm sequence. Because [Ger26] is the author's own companion paper and its statements are not reproduced in this preprint, the derivation chain is not self-contained at its most load-bearing point. Theorem 4.10 is likewise quoted from [Ger26] and feeds into Corollary 4.13 and Proposition 6.4. I do not raise the score to 6-8 because there is no exhibited constructional equivalence: the paper's theorem is not literally identical to its inputs, and the cited companion results could in principle be independently proved. If [Ger26] were included or independently machine-checked, the score would drop to 0-2. From the text alone, the central equivalence is supported by load-bearing self-citation, giving a moderate circularity score.

Assumptions & free parameters 0 free parameters · 7 assumptions · 3 invented entities

No numerical free parameters appear; this is a pure-math theorem paper. The real 'inputs pulled from upstream' are: (1) the restricted class of good adic spaces, explicitly acknowledged in §1.4; (2) the black-box import of the author's own companion paper [Ger26] for representability and higher direct images — the main circularity-adjacent burden; (3) a long list of cited external theorems (Scholze–Weinstein, Anschütz–Le Bras, Heuer, Scholze, Fargues–Scholze). The invented objects (relative Banach–Colmez spaces, analytic Dieudonné crystals) are recombinations of existing data and carry no independent empirical handles.

assumptions (7)
  • domain assumption The base S ranges over good adic spaces (Definition 2.4), i.e. sousperfectoid or rigid spaces with O_{S,ét} = ν_*O_{S,v}; all main theorems are stated only for this class.
    Restricts scope; §1.4 acknowledges the theory 'fundamentally relies on perfectoid methods' and that the diamond functor 'only remembers topological information'. If goodness fails (non-sheafy adic spaces or general v-stacks), the stated equivalences are not asserted.
  • ad hoc to paper The companion results of [Ger26] are correct: representability by analytic p-divisible groups ([Ger26, Lemma 3.38]), good-idadic-space framework ([Ger26, Prop. 2.4-2.5]), and diamantine higher direct images ([Ger26, Prop. 3.16, 3.39]).
    Invoked in the proofs of Theorem 3.13 (the 'G is representable' step), Section 2.4, Theorem 3.44 and Theorem 4.10 (quoted verbatim). [Ger26] is the author's own prior work (the author thanks his advisor A. Werner and 'Daniel Kim for pointing out an inaccuracy in an earlier version'), not included in this preprint; its assertions are load-bearing and unverifiable here.
  • standard math Scholze–Weinstein classification of p-divisible groups over O_C (Theorem 3.31, cited [SW13, Thm. B]) is correct and applies.
    External theorem, not proved here; used to prove that dualizable groups have good reduction v-locally (Prop. 3.34) and to compute height/dimension in Prop. 6.4(2).
  • standard math Relative Hodge–Tate spectral sequence and primitive comparison theorem for proper smooth rigid spaces (Theorems 2.5, 2.7, cited from [Heu25], [Sch13a]) are correct.
    Background from Heuer and Scholze; used to construct the Hodge–Tate sequence (2.8), Lemma 2.14, and the maps f_H in Section 4.
  • standard math The relative Fargues–Fontaine curve X_S, small v-stack descent for shtukas, Beauville–Laszlo glueing, and the classification of G-bundles are quoted from [SW20] and [FS24].
    Background for Section 5; Proposition 5.7 is reproduced from [SW20, Prop. 23.3.1].
  • standard math Prismatic Dieudonné theory of Anschütz–Le Bras [AL23] and log-prismatic Dieudonné theory [Ino25][WZ23] are proven equivalences.
    External theorems on which the compatibility results of Prop. 4.5, 4.7, 5.35 and 5.37 rest.
  • ad hoc to paper Example 3.32: there exists a formal elliptic curve E over Spf(C+) for a bounded open valuation subring C+ ⊆ C of finite Krull dimension ≥ 2 with E_{k(s)} supersingular and E_{k(u)} ordinary.
    Asserted with 'For example, s could be a higher rank point in the boundary of the supersingular locus inside the good reduction locus of the modular curve, viewed as an adic space'; no construction is given. Used only to illustrate failure of full faithfulness of (·)_η over arbitrary (C, C+), so it is illustrative rather than load-bearing for the main theorems.
invented entities (3)
  • Relative analytic p-divisible groups over good adic spaces (family version of Fargues' definition, Def. 3.1)
    purpose: Central object of study; the paper develops the family theory, the logarithm sequence, Weil pairings, and dualizability.
    The notion for points is due to Fargues [Far19]; the paper's contribution is the relative theory, not the entity itself. No empirical handle outside the mathematics.
  • Analytic Dieudonné crystal (Definition 5.25): minuscule shtuka whose φ(M)/M and dual quotient arise from étale vector bundles
    purpose: Target category of Dieudonné theory over arbitrary good adic spaces; Theorem 5.26 equates it with dualizable analytic p-divisible groups.
    New definition introduced here, recombining existing objects (minuscule shtukas + étale-bundle condition). Its justification is the equivalence theorem itself; no independent falsifiable handle.
  • Relative effective Banach–Colmez spaces (Definition 3.36)
    purpose: Classify p-adic universal covers eG of analytic p-divisible groups; presented effective Banach–Colmez spaces are classified by Theorem 3.44.
    Relative variant of Fontaine/Colmez spaces; new in families but built from Q_p-local systems and étale vector bundles. No independent empirical handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Dieudonn\'e theory for analytic p-divisible groups and applications to Shimura varieties." pith.science (2026). https://pith.science/paper/2QTM6V56

@misc{pith2026260205764,
  author       = {Pith},
  title        = {Pith review of: A Dieudonn\'e theory for analytic p-divisible groups and applications to Shimura varieties},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2QTM6V56}},
  note         = {Machine review of arXiv:2602.05764}
}
abstract

We study families of analytic $p$-divisible groups over adic spaces $S$ defined over $\mathbb{Q}_p$. We prove an equivalence between such families and Hodge-Tate triples, generalizing a theorem of Fargues. For a perfectoid space $S$, we construct a functor associating to an analytic $p$-divisible group $\mathcal{G} \rightarrow S$ a coherent sheaf $\mathcal{E}(\mathcal{G})$ on the relative Fargues--Fontaine curve $X_S$. Restricting to analytic $p$-divisible groups admitting a Cartier dual, we obtain an equivalence of categories with local shtukas satisfying a minuscule condition, compatible with the prismatic Dieudonn\'e theory of Ansch\"utz--Le Bras. We conclude with applications to moduli spaces: we show that the local Shimura varieties of EL and PEL types of Scholze--Weinstein are moduli spaces of analytic $p$-divisible groups with extra structure, and we give a reinterpretation of the Hodge--Tate period map of Scholze in terms of topologically $p$-torsion subgroups of abelian varieties.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

2 extracted references · 2 linked inside Pith

  1. [2]

    [Gle25] I

    arXiv:2309.06663 [math.AG]. [Gle25] I. Gleason. Specialization maps for Scholze’s category of diamonds.Math. Ann.391 (2), 1611–1679, 2025. [GR24] H. Guo and E. Reinecke. A prismatic approach to crystalline local systems.Invent. Math. 236 (1), 17–164, 2024. [Guo24] H. Guo. Crystalline cohomology of rigid analytic spaces. To appear inBull. Soc. Math. Fr., 2...

    arXiv 2025

  2. [2023]

    [Bos23b] G

    arXiv:2110.10683 [math.NT]. [Bos23b] G. Bosco. Rationalp-adic Hodge theory for rigid-analytic varieties. Preprint. 2023. arXiv: 2306.06100 [math.AG]. [BS17] B. Bhatt and P. Scholze. Projectivity of the Witt vector affine Grassmannian.Invent. Math.209 (2), 329–423, 2017. [BS22] B. Bhatt and P. Scholze. Prisms and prismatic cohomology.Ann. of Math.196 (3), ...

    arXiv 2023

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.