{"id":"c143e688-6977-4d3c-aa3c-a7bd8fe0d921","arxiv_id":"2602.05764","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Analytic p-divisible groups over good adic spaces are equivalent to Hodge-Tate triples (T_pG, Lie G, f_G), and dualizable ones to minuscule local shtukas — yielding moduli descriptions of EL/PEL local Shimura varieties.","lead":"This paper builds a family version of Dieudonné theory: analytic p-divisible groups over certain p-adic spaces are classified by linear-algebraic data, and — when they have a dual — by minuscule local shtukas. A generalist might care because it gives local Shimura varieties a moduli interpretation as families of analytic p-divisible groups, and reinterprets Scholze's Hodge–Tate period map in terms of topologically p-torsion subgroups of abelian varieties.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.13 rests on an unproved external lemma [Ger26, Lemma 3.38]; if that lemma fails in the stated generality, the paper's central equivalence collapses.","rationale":"The reader's weakest-assumption analysis already identified the correctness of the imported results from [Ger26] as the critical dependency. My stress-test confirms this is indeed the single most load-bearing concern. The proof of Theorem 3.13 explicitly invokes [Ger26, Lemma 3.38] at the two decisive steps: representability of the fiber product and identification of the logarithm sequence. Without that lemma, the equivalence is not established. The paper also quotes Theorem 4.10 from [Ger26] for higher direct images, which underpins later applications, but the primary dependence is on Lemma 3.38. I do not find an internal contradiction in the main argument when the imported results are assumed. The missing diagram in Proposition 5.37 and the duplicated sentence in §5.5 are presentation issues, not mathematical showstoppers. The unproved assertion in Remark 3.46(1) about isomorphisms of universal covers of non-isogenous abeloid varieties is also noteworthy, but it is used for future applications rather than for the central equivalence. Therefore, the verdict should remain CONDITIONAL, pending verification of [Ger26, Lemma 3.38] (and the related [Ger26] statements). My concrete test targets the minimal nontrivial instance of the lemma where the base is a non-perfectoid good adic space, because that is exactly the regime the paper claims to cover. If the lemma passes there, the central claim is substantially supported; if not, the theorem needs a different proof or a restriction on the base.","tokens_in":68901,"tokens_out":3082,"duration_ms":34049,"concrete_test":"Obtain [Ger26] and verify Lemma 3.38 in the full generality claimed. Specifically, check the special case where S = Spa(C, C^+) with C^+ a rank-2 valuation subring, L = Z_p, E = O_S, and f = 0: the lemma must yield an analytic p-divisible group (isomorphic to G_m⟨p^∞⟩) with the standard logarithm sequence. Also verify the same lemma over S = Spa(Q_p). If the lemma is valid only for perfectoid bases or under stronger assumptions on f, Theorem 3.13 requires reformulation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.13, the paper's central equivalence, is proved by constructing an inverse functor: a tuple (L,E,f) is sent to the fiber product of E⊗G_a and L(−1)⊗G_m⟨p^∞⟩ over L(−1)⊗G_a. The proof then asserts twice that [Ger26, Lemma 3.38] guarantees this fiber product is representable by an analytic p-divisible group and that its logarithm sequence is the expected one. Thus the entire classification of analytic p-divisible groups by Hodge–Tate triples depends on a lemma imported from a companion paper that is not included in this preprint and cannot be checked here. If [Ger26, Lemma 3.38] has hidden hypotheses—for example, if it requires the base to be perfectoid, or requires f to be a locally direct summand with vector-bundle cokernel—then Theorem 3.13, and with it all later results built on it (Theorems 5.23, 5.26, 6.7, 7.8), fail. The same external dependency appears in Theorem 4.10, quoted from [Ger26, Prop. 3.16, 3.39], which is used for Corollary 4.13 and Proposition 6.4. The manuscript does not reproduce these statements or proofs, so the validity of the central claim cannot be independently verified from the given text. This is a gap in support rather than an internal inconsistency, but it is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":69235,"tokens_out":9276,"duration_ms":109045,"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":[{"comment":"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.","section":"Theorem 3.13 (§3.3)"},{"comment":"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.","section":"Theorem 4.10 (§4.2)"},{"comment":"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.","section":"Theorem 5.26 and Definition 5.25 (§5.3)"}],"minor_comments":[{"comment":"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.","section":"§3.3, proof of Theorem 3.13"},{"comment":"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.","section":"§2.3, Lemma 2.15"},{"comment":"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.","section":"Example 3.32"},{"comment":"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.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper's central theorems are conditional on [Ger26] to an unusual degree: the main equivalence, the higher direct image machinery, and the descent argument all rely on lemmas not proved in this submission. I recommend that the editor require the author to either include the relevant proofs or post and cross-check the companion manuscript before acceptance. If the companion remains unavailable, the paper cannot be independently verified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a substantial paper and, if the companion paper holds, a major advance. The genuinely new content is the family version of Fargues' theorem (Thm 3.13), the relative sheaf E(G) on the Fargues–Fontaine curve (Thm 5.11), the analytic Dieudonné crystal equivalence over good adic bases (Thm 5.26), and the moduli reinterpretations of EL/PEL local Shimura varieties and the Hodge–Tate period map. The paper is honest about its limitations (§1.4), and the results are externally benchmarked against Fargues' pointwise theorem, Scholze–Weinstein, and Anschütz–Le Bras. That is real evidence that the architecture is not circular or self-serving.\n\nThe main soft spot is exactly where the reader's stress-test lands: the proof of Theorem 3.13, the load-bearing equivalence, relies on [Ger26, Lemma 3.38] for representability of the fiber product by an analytic p-divisible group and for the logarithm sequence. That lemma is not reproduced or proved here. The same companion paper supplies the higher-direct-image results (Theorem 4.10) and the good-idadic-space framework (Prop. 2.4–2.5). These are black boxes, and they are load-bearing. If any of them has hidden hypotheses—say, requiring a perfectoid base or stronger conditions on f—the central theorem and everything downstream collapses. I don't see an internal inconsistency, but I also cannot verify the paper's central claim from the text alone.\n\nThere are also smaller assembly problems: Proposition 5.37 has a missing diagram and a bare-reference proof, and §5.5 contains a duplicated sentence fragment. These are cosmetic but signal haste. Remark 3.46(1) asserts a non-trivial isomorphism of universal covers for non-isogenous abeloid varieties without proof, and then uses it to advertise future applications. That should be either proved or explicitly labeled as conjectural.\n\nProportionately: none of these are fatal in themselves, and the reader's conditional verdict is the right one. The paper deserves a serious referee, but only with [Ger26] in hand—either included as an appendix or made available. The comparison with Anschütz–Le Bras and the Scholze–Weinstein benchmarks suggest the main ideas are sound. I would send this to peer review, and I would specifically instruct the referee to check the external dependencies first.\n\nWho gets value: anyone working on p-divisible groups, local Shimura varieties, or relative p-adic Hodge theory. I would cite it once the companion is public, and I'd bring it to reading group now.","headline":"A genuinely strong family-level Dieudonné theory paper whose central theorem leans on unverified companion-paper lemmas; referee with [Ger26] in hand.","tokens_in":69840,"tokens_out":1077,"would_cite":true,"duration_ms":15802,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14L05","14G22","11G18"],"pacs":[],"model":"deepseek-v4-flash","headline":"Families of analytic p-divisible groups are classified by a single linear map f_G — this paper proves the equivalence.","keywords":["analytic p-divisible groups","Hodge–Tate triples","p-adic Hodge theory","Dieudonné theory","local shtukas","Shimura varieties","Cartier duality"],"falsifier":"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.","tokens_in":68561,"feed_emoji":"📐","tokens_out":9400,"duration_ms":99675,"temperature":0.7,"pith_summary":"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.","feed_headline":"One linear map governs analytic p-divisible groups","feed_subtitle":"The equivalence yields a Dieudonné theory and recasts local Shimura varieties as moduli of these groups.","key_machinery":"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","core_discovery":"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","pith_inferences":["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."],"forward_implications":["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."],"fun_headline_variants":["Analytic p-divisible groups yield Dieudonné theory","Local Shimura varieties as moduli of p-divisible groups","Hodge–Tate map gives equivalence for p-divisible groups","A triple equivalence for analytic p-divisible groups"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Analytic p-divisible groups yield Dieudonné theory","Local Shimura varieties as moduli of p-divisible groups","Hodge–Tate map gives equivalence for p-divisible groups","A triple equivalence for analytic p-divisible groups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000603,"raw_usage":{"total_tokens":2683,"prompt_tokens":805,"completion_tokens":1878,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":549,"completion_tokens_details":{"reasoning_tokens":1809}},"tokens_in":549,"tokens_out":1878,"duration_ms":13986,"temperature":1.0,"reasoning_tokens":1809,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T06:07:56.368998+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}