{"id":"0fe101a6-37d8-4585-80cf-5fcb32848ee5","arxiv_id":"2504.16282","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A functor D_crys is constructed that links prismatic F-crystals to filtered Frobenius crystals and, in the Fontaine-Laffaille range, induces an equivalence of categories.","lead":"This paper builds a new bridge between two modern frameworks in p-adic geometry, called prismatic cohomology and Fontaine-Laffaille theory. The bridge provides an integral version of a classical comparison functor, which helps connect different ways of classifying p-divisible groups.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The main equivalence hinges on Proposition 3.3, whose proof invokes a non-peer-reviewed lemma from [Hok24] and [Gao19, Lemma 3.8] as black boxes; a failure or misapplication there would break Theorem 3.1.","rationale":"The reader's conditional verdict is appropriate. I read Theorem 3.1, Proposition 3.4, and Proposition 3.3. The paper's own structure makes Proposition 3.3 load-bearing: Proposition 3.4 explicitly says Propositions 3.2 and 3.3 imply Theorem 3.1, and Proposition 3.3 is also used in Theorem 4.8 and Corollary 4.18. The rest of Section 2 is a detailed construction with a clean internal comparison theorem; if one accepts the cited foundations (GR24, DLMS24, IKY25, Tsu20, Fal89), the main new content is D_crys and the reduction. The only place where the argument outsources the key freeness assertion is the proof of Proposition 3.3. The reliance on [Hok24, Lemma 52] is acknowledged in Remark 1, but Remark 1 also claims independence from Hok24's results; that claim is weakened by the later use of Lemma 52. This is not a logical inconsistency, just an unverified dependency. No ad hominem: Hok24 may well be correct. But for a central equivalence theorem, an unreviewed preprint lemma plus an unverified applicability of Gao's lemma warrants conditional rather than full acceptance. The proposed test—an independent proof of the cokernel freeness in the power-series case—would settle it. I therefore keep the reader's conditional verdict unchanged.","tokens_in":37982,"tokens_out":9873,"duration_ms":91543,"concrete_test":"Prove [Hok24, Lemma 52] from first principles in the exact setting of Proposition 3.3: let R = W[[t_1,...,t_d]], S_R = R[[u]] with E = u, and let M be a finite projective S_R-module of height ≤ p−2 with φ_M(φ^*M) ⊆ E^{p−2}M. Show directly that coker(φ_M : φ^*M → M/E^i) is finite free over R for each i. A tractable first case is d = 1, rank M = 1, which can be checked by explicit matrices. In parallel, verify that the base-changed M_0 satisfies the hypotheses of [Gao19, Lemma 3.8] (not merely effectiveness), and that the total rank Σ_i r_i equals rank_R(φ^*M/E); if either fails, Proposition 3.4's reduction collapses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.1 is reduced in Proposition 3.4 to Propositions 3.2 and 3.3. Proposition 3.3 is the point where the argument is thinnest: for R = W[[t_1,...,t_d]], it needs the cokernel of f_i : φ^*_{S_R}M → M/E^i to be finite free over R, attributing this to 'the argument in [Hok24, Lemma 52]', and then needs [Gao19, Lemma 3.8] to know the graded pieces of the base-changed filtration are finite free over W(k). Both lemmas are imported as black boxes; [Hok24] is an unreviewed preprint (arXiv:2407.21327), and Remark 1 says only that the proof of Proposition 3.3 is 'inspired by' it, not that it is independent. If Hok24's lemma has hidden hypotheses (e.g., a different Frobenius, strongly divisible input, or an algebraically closed residue field) or if [Gao19, Lemma 3.8] is applied to M_0 before M_0 is known to satisfy its hypotheses, then the freeness of Gr^i(Fil^bar) is not established. Since Proposition 3.3 is exactly what upgrades the partially known arrows in diagram (3.1.1) to equivalences in the Fontaine–Laffaille range, the main theorem inherits this uncertainty.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs an integral analogue D_crys of Fontaine's crystalline functor for prismatic F-crystals and prismatic F-gauges on a smooth formal scheme over W(k), landing in categories of filtered F-crystals. It proves a crystalline--de Rham comparison for prismatic F-crystals (Theorem 1.19), compares D_crys with the rational crystalline functor (Theorem 2.10), characterizes local filtered freeness through D_crys (Proposition 2.13), and proves bi-exactness of the forgetful functor from prismatic F-gauges to lff prismatic F-crystals (Proposition 2.17). The central result is Theorem 3.1, which asserts that in the Fontaine--Laffaille range [0,p-2] the diagram relating prismatic F-crystals, prismatic F-gauges, local systems, and filtered F-crystals is 2-commutative and every arrow is a Z_p-linear equivalence. The paper also applies D_crys to prismatic Dieudonné theory, proving a filtered identification D ≃ D_crys ∘ M_Δ and relating the prismatic Dieudonné functor to Breuil--Kisin--Kim Dieudonné functors.","tokens_in":38280,"tokens_out":8704,"duration_ms":90843,"significance":"If the main theorem holds, the paper gives a genuinely useful bridge among the crystalline, prismatic, and syntomic approaches in the Fontaine--Laffaille range, and it substantially clarifies the integral relation between prismatic Dieudonné theory and classical crystalline Dieudonné theory. The paper is carefully organized and contains substantial proofs from definitions, especially for the crystalline--de Rham comparison (Theorem 1.19), the comparison with rational D_crys (Theorem 2.10), the lff characterization (Proposition 2.13), and the Dieudonné-theoretic applications (Theorem 4.8 and Corollary 4.18). The main reservation is that the proof of the crucial Proposition 3.3 delegates a key freeness assertion to an unpublished preprint and applies another external lemma without verifying its hypotheses; since Proposition 3.3 is what upgrades all other known equivalences to the full statement of Theorem 3.1, the central claim is conditional on that step.","major_comments":[{"comment":"The proof of Proposition 3.3 invokes 'the argument in [Hok24, Lemma 52]' to show that coker(f_i) is a finite free R-module for f_i : φ^*_{S_R}M → M/E^i. This is a load-bearing assertion: Proposition 3.3 is exactly the statement used in Proposition 3.4 to turn fully faithful or partially known arrows in diagram (3.1.1) into equivalences in the Fontaine--Laffaille range. However, [Hok24] is an unreviewed preprint (arXiv:2407.21327), and the manuscript does not reproduce the argument or state the precise hypotheses under which the lemma applies. The statement in Remark 1 that the proof is 'independent' of [Hok24] does not resolve this, because the proof as written uses that preprint as a black box. Please either include a complete proof of the cokernel freeness assertion or restate and prove the needed lemma in this paper, and verify explicitly that the setting of Proposition 3.3 satisfies all of its hypotheses.","section":"§3.4, proof of Proposition 3.3"},{"comment":"After reducing to R = WJt_1,...,t_dK, the proof applies [Gao19, Lemma 3.8] to the base-changed Breuil--Kisin module M_0 := M⊗_{S_R} S_W with its induced filtration, and uses the conclusion that Gr^i(Fil_0) is finite free over W(k) as the input for a Nakayama lifting argument. The manuscript does not verify that M_0 satisfies the hypotheses of [Gao19, Lemma 3.8]; in particular, it does not check the relevant strong divisibility or filtered-freeness conditions before applying the lemma. This matters because the freeness of Gr^i(Fil_0) is what produces the surjection R^{r_i} → Gr^i(Fil^bar) and ultimately the desired free basis of the graded pieces of Fil^bar. Please add the missing verification, or give a reference that contains it, so that the application of [Gao19, Lemma 3.8] is fully justified.","section":"§3.4, use of [Gao19, Lemma 3.8]"}],"minor_comments":[{"comment":"The word 'Dieduonné' should be 'Dieudonné' in the abstract and in the Section 4 heading; there are also typographical slips 'crysal' in §1.3 and 'siomorphism' in §2.1.3.","section":"Abstract and §4 heading"},{"comment":"The target of the composite functor in display (1) is written 'V ectφ,div(Xcrys)', but elsewhere, including (3.1.2), the same functor is said to take values in 'V ectFφ,div(Xcrys)'; the notation should be made consistent.","section":"Introduction, display (1)"},{"comment":"In the list of notation, the entry 'V ectan,φφ(X∆)' appears to be a typo for 'V ectφ,an(X∆)', and should be corrected.","section":"Notation and terminology"},{"comment":"The one-sentence reduction to R = WJt_1,...,t_dK should be expanded: after localizing at a maximal ideal of R/p, the residue field may be a finite extension of k rather than k itself, so the coefficient ring is a Witt ring W(κ) with κ≠k; please spell out the Cohen structure theorem step and how flat base change handles the residue-field extension.","section":"§3.4, reduction to power-series case"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi [Name],\n\nQuick take on Imai–Kato–Youcis, arXiv:2504.16282. The paper does what it says: it constructs an integral D_crys from prismatic F-crystals to filtered F-crystals, proves a crystalline–de Rham comparison, and shows that in the Fontaine–Laffaille range the whole diagram of prismatic, syntomic, and crystalline functors collapses to equivalences. That is a real and significant step, and the Dieudonné applications are a nice payoff.\n\nThe paper is well organized and the main new proofs are done in detail. I particularly like Proposition 2.13, where lffness is characterized by strong divisibility of D_crys, and Proposition 2.17 on bi-exactness. The comparison with Tsuji's functors in Section 3 is careful.\n\nThe soft spot is exactly where you'd expect from the stress-test: Proposition 3.3. The proof reduces to a power series ring, then says 'the argument in [Hok24, Lemma 52]' gives a finite free cokernel, and applies [Gao19, Lemma 3.8] to get freeness of the graded pieces. Hok24 is an unreviewed arXiv preprint, and this is a load-bearing step: without Proposition 3.3, the FL equivalence doesn't go through. It may well be true—the lemma is plausible and the reduction to the power-series case is standard—but importing it as a black box from an unreviewed source is a real weakness. I'd want a referee to verify that lemma, or the authors to supply a proof, before I'd fully trust Theorem 3.1. The dependency on Bha23 course notes is less worrying, since those are widely used.\n\nThe citation pattern looks honest; they explicitly compare with Hokaj, Würthen, and TVX24 and say where they overlap. The main theorem is not a repackaging of known results.\n\nVerdict: this deserves a serious referee. I'd recommend accepting for review, with the expectation that Prop 3.3 gets fixed or explicitly verified. For my own work, I'd likely cite it once the external dependency is cleaned up; I'd bring it to reading group in the meantime.\n\nBest,\n[Your name]","headline":"A substantial and well-written construction of an integral D_crys that completes the comparison of crystalline, prismatic, and syntomic approaches, but with a load-bearing step in Proposition 3.3 that imports a lemma from an unreviewed preprint as a black box.","tokens_in":38833,"tokens_out":2673,"would_cite":true,"duration_ms":24678,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F30","14L15","11F80"],"pacs":[],"model":"deepseek-v4-flash","headline":"Integral p-adic Hodge theories unify in the Fontaine–Laffaille range","keywords":["prismatic F-crystals","Fontaine–Laffaille modules","filtered F-crystals","crystalline-de Rham comparison","Nygaard filtration","p-divisible groups","Dieudonné theory","integral p-adic Hodge theory"],"falsifier":"Look for a prismatic F-crystal of height at most $p-2$ over $R=W[[t_1,\\dots,t_d]]$ for which some graded piece $\\mathrm{Gr}^i(\\mathrm{Fil}^\\bullet)$ of the Nygaard filtration on $\\varphi^*E/(E)$ is not finite free; Proposition 3.3 asserts none exists, so one such example would falsify the main equivalence. Concretely, compute the cokernel of the relative Frobenius map $\\varphi_M:\\varphi^*M\\to M/E^i$ for an explicit Breuil–Kisin module and check whether the quoted Lemma 52 of the cited preprint holds there.","tokens_in":2328,"feed_emoji":"","tokens_out":3520,"duration_ms":122164,"temperature":0.7,"pith_summary":"The paper constructs a functor $D_{\\mathrm{crys}}$ from prismatic F-crystals and prismatic F-gauges on a smooth formal scheme over the Witt vectors into filtered F-crystals, and proves it behaves like an integral version of Fontaine's crystalline functor: after inverting $p$ it matches the rational functor via the étale realization. Its main theorem, for $p>2$, is a 2-commutative diagram in which prismatic F-crystals, prismatic F-gauges, crystalline local systems, and Fontaine–Laffaille modules with Hodge–Tate weights in $[0,p-2]$ are all $\\mathbb{Z}_p$-linearly equivalent through $D_{\\mathrm{crys}}$, the forgetful functor, and the two realization functors. It also shows that the prismatic Dieudonné crystal of a $p$-divisible group recovers the filtered crystalline Dieudonné crystal, and that the Breuil–Kisin–Kim Dieudonné functor is evaluation of the prismatic Dieudonné crystal at the Breuil–Kisin prism. The argument rests on a canonical comparison between crystalline and de Rham realizations, and on a reduction to power-series rings where the lowness of the Hodge–Tate weights forces the relevant filtered pieces to be free.","feed_headline":"Integral p-adic Hodge theories unify in the Fontaine–Laffaille range","feed_subtitle":"New functor Dcrys makes prismatic F-gauges and Fontaine–Laffaille modules equivalent, and recovers filtered Dieudonné crystals.","key_machinery":"The load-bearing object is the functor $D_{\\mathrm{crys}}$, defined as follows: take a prismatic F-crystal $(E,\\varphi_E)$, form its crystalline realization $E_{\\mathrm{crys}}$ and its de Rham realization $E_{\\mathrm{dR}}$, and use the canonical identification $\\iota_X$ between them to transfer the Nygaard filtration of $\\varphi^*E$ onto $E_{\\mathrm{crys}}$; the resulting filtered F-crystal is $D_{\\mathrm{crys}}(E,\\varphi_E)$. Local filtered freeness of the Nygaard filtration is the condition that makes $D_{\\mathrm{crys}}$ land in strongly divisible filtered F-crystals and lets it detect lffness and control exactness. In the proof of the Fontaine–Laffaille equivalence, the argument reduces to power-series rings and uses a lemma asserting that the relative Frobenius map on the associated Breuil–Kisin module has finite free cokernel, together with a finiteness result for the graded pieces over $W$.","core_discovery":"The central claim is that the crystalline, prismatic, and syntomic routes to integral $p$-adic Hodge theory carry equivalent information inside the Fontaine–Laffaille range. For a smooth formal $W$-scheme $X$ and $p>2$, the paper defines $D_{\\mathrm{crys}}(E,\\varphi_E)$ using the crystalline realization $E_{\\mathrm{crys}}$ and the de Rham realization $E_{\\mathrm{dR}}$, joined by a canonical isomorphism $\\iota_X$ (Theorem 1.19); the filtration is pulled back from the Nygaard filtration on $\\varphi^*E$. Theorem 3.1 then states that the diagram linking locally filtered free prismatic F-crystals, prismatic F-gauges, all prismatic F-crystals in the range, crystalline local systems, and strongly divisible filtered F-crystals commutes and every arrow is a $\\mathbb{Z}_p$-linear equivalence. Theorem 4.8 upgrades the earlier prismatic Dieudonné comparison to a filtered one: $D_{\\mathrm{crys}}\\circ M_\\Delta$ is naturally the filtered crystalline Dieudonné functor $D$, and Proposition 4.16 identifies the Breuil–Kisin–Kim functor with evaluation of $M_\\Delta$ on the Breuil–Kisin prism.","pith_inferences":["Beyond the paper, the same comparison should make it possible to transport quantitative Fontaine–Laffaille estimates, such as bounds on extension groups or filtration jumps, into the prismatic category, where such bounds are harder to see; one test would be to reprove the classical strong-divisibility results directly through $D_{\\mathrm{crys}}$.","Beyond the paper, the construction suggests a derived or $\\infty$-categorical analogue of $D_{\\mathrm{crys}}$ for perfect complexes on the prismatic site, since the crystalline–de Rham comparison is naturally formulated with perfect complexes and the authors note the isomorphism should persist after inverting $p$.","Beyond the paper, the explicit comparison on Breuil–Kisin prisms provides a coordinate formula for $D_{\\mathrm{crys}}$, so a reader could use it as a computational tool to compute filtrations attached to explicit prismatic F-crystals and $p$-adic representations.","Beyond the paper, bi-exactness of the forgetful functor indicates that the prismatic and syntomic sites compute the same extension groups; a concrete check is to compare $\\mathrm{Ext}^1$ in both categories for small $p$-divisible groups."],"forward_implications":["In Hodge–Tate weights $[0,p-2]$, the prismatic, syntomic, and crystalline categories are equivalent, so a crystalline local system admits matching integral models and one can pass freely between the three frameworks.","$D_{\\mathrm{crys}}$ gives a canonical filtered lattice inside the rational crystalline functor, so the filtration recorded by Grothendieck–Messing theory is determined by the prismatic Dieudonné crystal, not just its underlying F-crystal.","The forgetful functor from prismatic F-gauges to locally filtered free prismatic F-crystals is bi-exact, which controls extension groups in both categories.","For smooth proper $X$, the comparison yields an isomorphism between étale cohomology of $T_{\\acute{e}t}(V)_{\\mathrm{alg}}$ modulo $p^n$ and $T_{\\acute{e}t}(R^b f_* V/p^n)$ whenever $a+b<p-2$.","For every formally framed $W$-algebra $R$, the categories $Vect_{[0,1]}(R_{\\mathrm{syn}})$, $Vect^\\varphi_{[0,1]}(R_\\Delta)$, and $Vect^\\varphi_{[0,1]}(S_R,\\nabla^0)$ are equivalent, identifying prismatic Dieudonné theory with Breuil–Kisin–Kim Dieudonné theory."],"supporting_citations":[{"why":"Supplies the Fontaine–Laffaille category, the crystalline realization functor $T_{\\mathrm{crys}}$, and the strong-divisibility framework that the paper's $D_{\\mathrm{crys}}$ targets; its Theorem 2.6 gives the full faithfulness used in Theorem 3.1.","marker":"[Fal89]"},{"why":"Introduced prismatic F-crystals and the crystalline realization $E_{\\mathrm{crys}}$ that forms the underlying piece of $D_{\\mathrm{crys}}$; Example 4.12 underlies Section 1.1.","marker":"[BS23]"},{"why":"Provides the quoted Lemma 52, used in Proposition 3.3, asserting that the relative Frobenius map on the Breuil–Kisin module has finite free cokernel after reduction to the power-series case.","marker":"[Hok24]"},{"why":"Provides Lemma 3.8, used in Proposition 3.3 to know that the graded pieces of the filtered Breuil–Kisin module over $W$ are finite free.","marker":"[Gao19]"},{"why":"Establishes full faithfulness of the étale realization $T_{\\acute{e}t}$, a key arrow in the equivalence statement of Theorem 3.1.","marker":"[GR24]"},{"why":"Supplies the theory of prismatic F-gauges and the forgetful functor $R_X$ that connects the syntomic and prismatic categories in the main diagram.","marker":"[GL23]"},{"why":"Constructs the prismatic Dieudonné functor $M_\\Delta$ whose comparison with the filtered crystalline Dieudonné functor is Theorem 4.8, and gives the evaluation identification used at $R=W$.","marker":"[ALB23]"},{"why":"Constructs the Breuil–Kisin–Kim Dieudonné functor $M$ and the category $Vect^\\varphi_{[0,1]}(S_R,\\nabla^0)$ that Proposition 4.16 ties to $M_\\Delta$.","marker":"[Kim15]"},{"why":"Provides the auxiliary functors $T_{A_{\\mathrm{inf}}}$ and $T_{A_{\\mathrm{crys}}}$ and Theorem 63(2), used in the proof of Proposition 3.2 to match $T_{\\mathrm{crys}}\\circ D_{\\mathrm{crys}}$ with $T_{\\acute{e}t}$.","marker":"[Tsu20]"}],"fun_headline_variants":["Prismatic F-gauges now equivalent to Fontaine–Laffaille modules","New functor Dcrys recovers filtered Dieudonné crystals","Integral p-adic Hodge theory unified in low weights","Prismatics and Fontaine–Laffaille align via Dcrys"],"cache_read_input_tokens":40960,"weakest_assumption_plain":"The proof of the key Proposition 3.3 assumes on the authority of an unreviewed preprint that a relative Frobenius map on the relevant Breuil–Kisin module has finite free cokernel, so the Fontaine–Laffaille-range equivalence would collapse if that lemma were false.","fun_headline_variants_meta":{"raw":{"variants":["Prismatic F-gauges now equivalent to Fontaine–Laffaille modules","New functor Dcrys recovers filtered Dieudonné crystals","Integral p-adic Hodge theory unified in low weights","Prismatics and Fontaine–Laffaille align via Dcrys"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000993,"raw_usage":{"total_tokens":4289,"prompt_tokens":1110,"completion_tokens":3179,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":726,"completion_tokens_details":{"reasoning_tokens":3100}},"tokens_in":726,"tokens_out":3179,"duration_ms":23956,"temperature":1.0,"reasoning_tokens":3100,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:06:31.302277+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for a prismatic F-crystal of height at most $p-2$ over $R=W[[t_1,\\dots,t_d]]$ for which some graded piece $\\mathrm{Gr}^i(\\mathrm{Fil}^\\bullet)$ of the Nygaard filtration on $\\varphi^*E/(E)$ is not finite free; Proposition 3.3 asserts none exists, so one such example would falsify the main equivalence. Concretely, compute the cokernel of the relative Frobenius map $\\varphi_M:\\varphi^*M\\to M/E^i$ for an explicit Breuil–Kisin module and check whether the quoted Lemma 52 of the cited preprint holds there.","supporting_citations":[],"review_version":1}