REVIEW 3 major objections 4 minor 18 references
Prismatic cohomology and $A_{\inf}$-cohomology with coefficients
T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read Theorem 0.1: for a smooth p-adic formal scheme over a perfectoid integer ring, the prismatic cohomology of a locally finite free prismatic crystal is canonically isomorphic to the A_inf-cohomology of the corresponding relative Breuil-Kisin-
desk verdict Serious and credible extension of the prismatic / A_inf comparison to locally finite free coefficients; main risk is the paper's reliance on the companion [17] and the non-cartesian simplicial descent in §11, not the local comparison. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The argument is carried by three objects working together. (1) The q-Higgs module formalism: on a framed smooth q-prism (D, t, θ_D), a prismatic crystal corresponds to an integrable connection over twisted derivations θ_{D,i}, called a q-Higgs field; its de Rham complex, the q-Higgs complex qΩ•(F_D), computes the prismatic cohomology of F in the affine case (Theorems 4.13 and 4.20). (2) The relative Breuil-Kisin-Fargues module M_{BKF,X}(F), a locally finite free A_inf-module on the proétale site that is trivial modulo μ, with AΩ_X(M) = Lη_μ Rν_{X,*} M. (3) Cohomological descent: the global description of R u_{X/A_inf,*}F via q-Higgs complexes on a cosimplicial envelope of an admissible frame
What would settle it
Work out both sides of (0.2) for a genuinely non-constant crystal on a smooth affine formal scheme, e.g., a rank-one crystal on the formal affine line with a specified q-Higgs field: write the q-Higgs complex qΩ•(F_D), form the η_μ-truncated proétale complex K•_Λ(ι^*ν^∞_*M), and check that the comparison map is an isomorphism on every cohomology sheaf; a single nonzero class in a kernel or cokernel annihilated by Ker(A_inf → W(k)) on one side but not the other would refute the local case, and hence the global theorem to which it reduces.
Extended reading notes
Core claim
The central claim (Theorem 0.1, stated as Theorem 10.1) is that for a quasi-compact, separated, smooth p-adic formal scheme X over the ring of integers O of a perfectoid field C of mixed characteristic (0,p) containing all p-power roots of unity, and for any locally finite free crystal F on the prismatic site (X/(A_inf,(ξ)))_Δ, there is a canonical isomorphism in D⁺(X_Zar, A_inf), functorial in F: R u_{X/A_inf,*} F ≅ AΩ_X(M_{BKF,X}(F)). The right-hand side is the A_inf-cohomology of the relative Breuil-Kisin-Fargues module M assigned to F, defined as Lη_μ of the derived pushforward of M to the Zariski site. The comparison map is constructed one step at a time: locally on framed small affine
Load-bearing premise
The global comparison is assembled from a black-box description of prismatic cohomology by q-Higgs complexes over an admissible framed embedding system, taken from the companion paper: if that descent description, or the treatment of non-cartesian direct images of simplicial topos, fails, the local framed isomorphisms do not glue into Theorem 0.1.
Editorial extensions
If this is right
- The two integral coefficient theories — prismatic crystals and relative Breuil-Kisin-Fargues modules — produce canonically identical cohomology, so computations and structures from either side transfer.
- For a crystal equipped with a Frobenius structure, inverting μ yields an isomorphism (R u_{X/A_inf,*}F)[1/μ] ≅ Rν_{X,*}(L ⊗ A_inf,X)[1/μ], giving a distinguished triangle relating the étale cohomology of the lisse Z_p-sheaf L to the prismatic cohomology of F (0.5).
- When X is proper over O and C is algebraically closed, the comparison yields RΓ((X/(A_inf,(ξ)))_{Δ}, F)[1/μ] ≅ RΓ(X_proét, L) ⊗⁽L_{Z_p} A_inf[1/μ], a form of primitive comparison with coefficients.
- The compatibility with pullbacks, Frobenius, and tensor products makes the isomorphism a functorial identification of both theories as sheaves of A_inf-modules on the Zariski site, compatible with the ring structure where applicable.
Reading between the lines
- The q-Higgs descent description is a stronger, more computable object than either cohomology theory alone; it should make explicit calculations feasible on simplicial charts and may carry the full A_inf-cohomology ring structure once combined with the paper's tensor-product compatibility.
- The local-to-global strategy suggests a route to relative versions of the theorem for a smooth morphism X → Y, a direction the paper itself poses as a natural question by analogy with known constant-coefficient relative results.
- The machinery for sheaves with action of a profinite group and the Koszul resolutions computing the group cohomology of Γ_Λ = Map(Λ, Z_p) is developed independently of the main comparison and could be reused for neighbouring coefficient theories such as syntomic or log-prismatic cohomology.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves a comparison theorem (Theorem 0.1 / Theorem 10.1) between prismatic cohomology with locally finite free crystal coefficients and A_inf-cohomology with relative Breuil--Kisin--Fargues module coefficients, for quasi-compact separated smooth p-adic formal schemes over the ring of integers of a perfectoid field containing all p-power roots of unity. The proof proceeds by a local framed comparison via q-Higgs complexes, then a global cohomological-descent argument. The paper also establishes functoriality, compatibility with Frobenius pullback, and compatibility with tensor products.
Significance. If the proof is complete, this is a substantial result: it gives a canonical integral comparison of two cohomology theories with arbitrary locally finite free coefficients, extending Bhatt--Scholze's constant-coefficient comparison and making precise the role of relative Breuil--Kisin--Fargues modules. The local comparison is explicit and the compatibility statements are valuable. A notable strength is that the construction of the comparison map is given in detail in the framed affine case, with an explicit Koszul/q-Higgs description. However, the global theorem depends crucially on results quoted from the author's companion work [17] and on a non-cartesian simplicial descent formalism that is not visible in the available text; without those, the central claim is only conditional.
major comments (3)
- [§4, Theorems 4.13, 4.20, 4.26; Remark 4.25; formula (4.32)] The local and global descriptions of prismatic cohomology in terms of q-Higgs modules and cohomological descent are quoted from the companion paper [17] without proof. Since these are the structural input for the entire comparison—Theorem 4.26 gives the simplicial version and (4.32) is the descent formula used in the global argument—the paper's main theorem is conditional on the validity of [17]. The author should either include full proofs of these theorems, or state clearly and precisely which results are imported and provide enough detail for the referee to verify their hypotheses are satisfied in the present setting. As it stands, this is a load-bearing gap in the manuscript.
- [Introduction; §11; Theorem 4.26; (4.32)] The global passage from the local framed case to arbitrary X uses a direct image functor between simplicial toposes that the author explicitly says is not cartesian and is treated in §11. In the available text, §11 is not included; only its title appears in the table of contents. The canonicality of the simple complex in (4.32) and the coherence of the non-cartesian simplicial direct image Rν_{X_•,t_•,*} are therefore not verifiable from the manuscript. Unless §11 is supplied and shown to satisfy the simplicial identities needed for the Cech-descent comparison, the global isomorphism does not follow from the local Theorem 9.1.
- [Theorem 10.1 / §10] The main theorem is stated as Theorem 0.1 in the Introduction and referenced as Theorem 10.1, but the text of §10 is not present in the submitted version. The global comparison argument, including the role of admissible framed embedding systems and the descent along the simplicial resolution, cannot be assessed. The manuscript should include the complete Section 10, not merely its statement.
minor comments (4)
- [Throughout] The notation 'A inf' and 'A_inf' is used inconsistently; it should be typeset uniformly as A_inf. Similarly, 'pro´ etale' appears with a corrupted accent in several places.
- [Introduction, Theorem 0.1] The numbering of the main theorem in the Introduction and in Section 10 should be aligned, and the cross-reference should be explicit when the theorem is first stated.
- [§8, diagram (8.26)] The diagram (8.26) has several unlabeled arrows and the middle vertical map is not explicitly named; adding names would improve readability.
- [§9, Lemma 9.16] The phrase 'certain numbers of copies' is vague. Even if the count is irrelevant for the proof, stating the exact multiplicities or at least a formula would make the argument easier to follow.
Circularity Check
No significant circularity: the comparison is between two independently defined cohomology theories; heavy reliance on the author's prior work is reliance on independent prior theorems, not circular input.
full rationale
Theorem 0.1 compares the prismatic cohomology Ru_{X/A_inf,*}F, defined via the prismatic site, with the A_inf-cohomology AOmega_X(M) of the associated relative Breuil-Kisin-Fargues module M = M_BKF,X(F). These two objects are defined independently: M_BKF,X(F) is constructed by evaluating the prismatic crystal on affinoid perfectoids (Definition 6.5), and AOmega_X(M) is defined as L eta_mu R nu_{X,*} Rlim M (Definition 6.21). No parameter is fitted to a subset of the target data and then renamed a prediction. The proof uses Theorems 4.13, 4.20, and 4.26 from the companion paper [17] to describe the prismatic side in terms of q-Higgs complexes and cohomological descent; these are stated as prior theorems with explicit assumptions (admissible framed smooth q-pairs, q-prisms, etc.) and do not assume the isomorphism being proved. Remark 4.25, also from [17, 15.1], supplies the existence of admissible framed embedding systems; this is an independent geometric input. The non-cartesian simplicial topos direct image discussed in the Introduction and Section 11 is a technical tool developed in this paper; if its coherence properties failed, the global descent step would not go through, but that would be a mathematical gap or correctness risk, not circularity. The paper also situates itself against external benchmarks such as Bhatt-Scholze's constant-coefficient case and other prior comparisons (Remark 0.3), confirming that the general statement has independent content. Accordingly, no self-definitional, fitted-input, or self-citation chain reduces the central claim to its own inputs.
Assumptions & free parameters
assumptions (4)
- domain assumption The author's prior work [17] provides the q-Higgs module description of prismatic crystals and their cohomology (Theorems 4.13, 4.20, 4.26) and cohomological descent for Ru_{X/R,*}F in terms of q-Higgs complexes.
- domain assumption Every quasi-compact separated smooth p-adic formal scheme over a q-prism admits an admissible framed embedding system (Remark 4.25, [17, 15.1]).
- domain assumption Scholze's primitive comparison theorem [15, Theorem 5.1] and standard perfectoid facts used for structural properties of relative BKF modules in Section 6 (e.g., A_inf,X(V) congruent to A_inf(A_plus_V), discreteness results).
- standard math Set-theoretic conventions with two universes V in U for sites and topos (Introduction).
invented entities (1)
-
The morphism nu_{X,t} from the proetale topos to the topos of Gamma_Lambda-sheaves on X_Zar (Section 7)
Cite this review
Pith. "Pith review of Prismatic cohomology and $A_{\inf}$-cohomology with coefficients." pith.science (2026). https://pith.science/paper/4F7M5X42
@misc{pith2026250904954,
author = {Pith},
title = {Pith review of: Prismatic cohomology and $A_\inf$-cohomology with coefficients},
year = {2026},
howpublished = {\url{https://pith.science/paper/4F7M5X42}},
note = {Machine review of arXiv:2509.04954}
}
abstract
For a smooth $p$-adic formal scheme over the ring of integers of a perfectoid field of mixed characteristic $(0,p)$ containing all $p$-power roots of unity, we prove that the prismatic cohomology of a locally finite free prismatic crystal is isomorphic to the $A_{\inf}$-cohomology of the corresponding relative Breuil-Kisin-Fargues module, which is a certain type of locally finite free $\mathbb{A}_{\inf}$-module, on the pro\'etale site of the generic fiber. We use a global description of the former in terms of $q$-Higgs modules via cohomological descent. We also discuss its compatibility with inverse image functors, scalar extensions under Frobenius, and tensor products.
Reference graph
Works this paper leans on
-
[17]
Tsuji,Prismatic crystals andq-Higgs fields, arXiv:2403.11676v1 [math.AG]
T. Tsuji,Prismatic crystals andq-Higgs fields, arXiv:2403.11676v1 [math.AG]
-
[1]
Available at https://stacks.math.columbia.edu
The Stacks Project. Available at https://stacks.math.columbia.edu
- [2]
- [3]
-
[4]
B. Bhatt,Specializing varieties and their cohomology from characteristic 0 to characteristicp, Algebraic geometry: Salt Lake City 2015, Proc. Sympos. Pure Math., vol. 97.2, Amer. Math. Soc., 2018, pp. 43–88
work page 2015
- [5]
-
[6]
B. Bhatt and P. Scholze,Prisms and prismatic cohomology, Ann. of Math. (2)196(2022), no. 3, 1135–1275
work page 2022
-
[7]
Faltings,Ap-adic Simpson correspondence, Adv
G. Faltings,Ap-adic Simpson correspondence, Adv. Math.198(2005), 847–862
work page 2005
Show all 18 references
-
[8]
Gaisin and Koshikawa
I. Gaisin and Koshikawa. T.,RelativeA inf -cohomology, arXiv:2206.07983v1 [math.NT]
-
[9]
Grothendieck,Revˆ etements ´ etales et groupe fondamental (SGA1), Lecture Notes in Math., vol
A. Grothendieck,Revˆ etements ´ etales et groupe fondamental (SGA1), Lecture Notes in Math., vol. 224, Springer, 1971
1971
-
[10]
Guo and E
H. Guo and E. Reinecke,A prismatic approach to crystalline local systems, Invent. Math.236(2024), no. 1, 17–164
2024
-
[11]
Hartshorne,Residues and duality, Lecture Notes in Mathematics, vol
R. Hartshorne,Residues and duality, Lecture Notes in Mathematics, vol. 20, Springer, 1966
1966
-
[12]
Koshikawa and Z
T. Koshikawa and Z. Yao,Logarithmic prismatic cohomology II, arXiv:2306.00364v1 [math.AG]
-
[13]
Min and Y
Y. Min and Y. Wang,Relative(φ,Γ)-modules and prismaticF-crystals, Math. Z.310(2025), no. 1, Paper No. 16, 25pages
2025
-
[14]
Morrow and T
M. Morrow and T. Tsuji,Generalised representations as q-connections in integralp-adic Hodge theory, arXiv:2010.04059v2 [math.NT]
2010 arXiv
-
[15]
Scholze,p-adic Hodge theory for rigid-analytic varieties—corrigendum [MR3090230], Forum Math
P. Scholze,p-adic Hodge theory for rigid-analytic varieties—corrigendum [MR3090230], Forum Math. Pi 4(2016), e6, 4
2016
-
[16]
Tian,A prismatic-etale comparison theorem in the semistable case, arXiv:2507.08451v1 [math.AG]
Y. Tian,A prismatic-etale comparison theorem in the semistable case, arXiv:2507.08451v1 [math.AG]
-
[18]
Graduate School of Mathematical Sciences, The University of Tokyo, 3-8-1 Komaba, Meguro- ku, Tokyo, 153-8914, Japan Email address:t-tsuji@ms.u-tokyo.ac.jp 86
,CrystallineZ p-representations andA inf -representations with Frobenius,p-adic Hodge theory, Si- mons Symp., Springer, 2020, 161–319. Graduate School of Mathematical Sciences, The University of Tokyo, 3-8-1 Komaba, Meguro- ku, Tokyo, 153-8914, Japan Email address:t-tsuji@ms.u...
2020
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.