{"id":"671082ec-c207-4a42-9b28-2928ca02d9c3","arxiv_id":"1908.07185","paper_version":4,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Every mod p representation of the absolute Galois group of a p-adic local field lifts to a crystalline representation of regular Hodge-Tate weights, via the geometry of new moduli stacks of etale (phi,Gamma)-modules.","lead":"Emerton and Gee build moduli stacks whose points are families of (phi,Gamma)-modules, algebraic objects that package the local Galois representations of a p-adic field. Using the geometry of these stacks they prove that every mod p Galois representation admits a lift to characteristic zero that is crystalline.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix F's 'admits all descents' criterion is the load-bearing premise; no internal flaw seen, but a rank-one explicit check would settle it.","rationale":"The reader's weakest_assumption is exactly mine: Definition 4.2.4 and Appendix F. I found no internal inconsistency in the sections reviewed; the proof of Proposition 1.5.1 is a disclosed bootstrap (Remark 1.5.4), not a circle, and the Noetherianity/dimension results have a detailed stack-theoretic apparatus. The reason I do not raise the equidimensionality induction as an alternative is that the authors explicitly flag the dependency: the dimension upper bound and the H^2-codimension estimate in Theorem 5.5.11 are independent of the crystalline substacks, while the full equidimensionality is deduced afterwards from Theorem 1.2.2. Thus the only place where a single technical error would propagate to the main theorem is the identification of the crystalline locus by BKF modules. Since Appendix F was truncated from the reviewed text, I cannot check the proof line-by-line; the proposed rank-one test uses established Fontaine/Kisin results and would mechanically verify the two directions of the characterization in the simplest nontrivial case. If the test passes, the accept verdict stands; if it fails, the verdict should move to conditional or reject. Because we currently have no evidence of failure, I keep the reader's verdict unchanged.","tokens_in":64497,"tokens_out":12463,"duration_ms":146629,"concrete_test":"Work out d=1 over a ramified K (e.g. K=Q_{p^2}) with the known crystalline character of a nonzero Hodge--Tate weight. Using only Fontaine/Kisin theory (not Appendix F), compute its BKF module M_inf and, for every uniformizer pi up to O_K^times-equivalence, the descent M_pi; verify M_pi subset (M_inf)^{G_{K_{pi,infinity}}}, the two independence conditions in Definition 4.2.4, and the inclusion (4.2.5). Then do the same for a semistable-but-not-crystalline character with the same Hodge type and check that (4.2.5) fails. If the crystalline character fails any condition, the Appendix F characterization is under-inclusive; if the semistable character satisfies all conditions, it is over-inclusive. This one computation would settle whether the concern lands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central lift theorem 1.2.2 goes through Theorem 1.2.4, whose crystalline substack is cut out by the new Breuil--Kisin--Fargues criterion Definition 4.2.4 ('admits all descents', proved in Appendix F). This is the true load-bearing seam: Theorem 1.2.4 is used in Section 1.5 to deduce effectivity of Spf R^{crys,lambda}/p -> X_d, and Corollary 1.5.2 then supplies the H^2-codimension bound that proves Theorem 1.2.2. If Definition 4.2.4 were under-inclusive (e.g. if a crystalline representation's BKF lattice is not G_{K_{pi,infinity}}-invariant for every uniformizer pi, or the inclusion (4.2.5) is too strong), the crystalline substack would be too small and the effectivity argument would collapse. If it were over-inclusive (e.g. a non-crystalline semistable representation also satisfies the two independence conditions (1)--(2) without satisfying (4.2.5)), Corollary 1.5.2 would not control H^2 on the true crystalline locus. The authors themselves remark in Section 1.6 that these compatibilities are likely better expressed via the prismatic site, which signals that the criterion is subtle. No internal inconsistency is apparent in the reviewed text, and the bootstrap in Remark 1.5.4 is not circular, but the Appendix F characterization is not independently supported by any machine-checked or previously established statement.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs, for a finite extension K of Q_p and a non-negative integer d, the stack X_d over Spf Z_p whose A-valued points, for p-adically complete Z_p-algebras A, are rank-d projective étale (phi,Gamma)-modules with A-coefficients. The main structural theorem (Theorem 1.2.1) states that X_d is a Noetherian formal algebraic stack whose reduced special fibre X_{d,red} is of finite type over F_p, equidimensional of dimension [K:Q_p]d(d-1)/2, and whose irreducible components admit a natural labelling by Serre weights. The paper also constructs closed crystalline and semistable substacks X_d^{crys,lambda} and X_d^{ss,lambda} (Theorem 1.2.4), using a new Breuil-Kisin-Fargues 'admits all descents' criterion (Definition 4.2.4, Appendix F). These stacks are used to prove that every continuous representation rho: G_K -> GL_d(F_p) admits a lift rho^circ to GL_d(Z_p) whose associated p-adic representation is crystalline of regular Hodge-Tate weights, and that the lift may be taken potentially diagonalizable (Theorem 1.2.2). Further consequences include a potential-automorphy globalization statement (Theorem 1.2.3), an explicit rank-one description of X_d (Section 7), and a geometric Breuil-Mezard conjecture (Conjecture 1.7.2, Section 8) with supporting results for GL_2.","tokens_in":64592,"tokens_out":9239,"duration_ms":99700,"significance":"If correct, this is a foundational contribution to the arithmetic of local Galois representations. It algebraizes Mazur's formal deformation rings and provides the first proof that arbitrary mod p representations of G_K admit crystalline lifts in all dimensions; as the authors explain in Section 1.5, this statement was not accessible to the standard inductive lifting of extension classes because of nonzero H^2 obstructions. The proof architecture is unusually transparent about dependencies: Remark 1.5.4 explicitly identifies the bootstrap by which Proposition 1.5.1, Corollary 1.5.2, and the effectivity of Spf R^{crys,lambda}/p -> X_d are used, and the apparent circularity is resolved rather than concealed. The geometric Breuil-Mezard statement is honestly labelled as Conjecture 1.7.2 and is not overclaimed. The paper also contains technical results of independent value, notably the faithfully flat descent theorem for projective modules over rings of Witt vectors of perfectoid fields (Theorem 1.6.1, proved as Theorem 2.4.1) and the new characterization of integral lattices in potentially semistable representations in Appendix F.","major_comments":[],"minor_comments":[{"comment":"The title displays 'ST ACKS' (an apparent spacing artifact) and the abstract contains several garbled accents such as '´ etale'; these should be corrected in the final version.","section":"Title and abstract"},{"comment":"In Theorem 1.2.2 the same symbol rho is used for the mod p representation and for its characteristic-zero lift before the notation rho^circ is introduced later in the paper; the introduction should use rho^circ consistently from the first statement.","section":"§1.2 and §6.4"},{"comment":"The term 'basic' is defined in §2.1.12 and subsequently used with two spellings ('K basic' and 'K_basic'); the notation should be unified, for example as K_basic throughout.","section":"§2.1.12 and §3.2.3"},{"comment":"Since the 'admits all descents' criterion is load-bearing for Theorem 1.2.4 and hence for Theorem 1.2.2, a short remark explaining the intended meaning of the two independence conditions (1) and (2), together with an explicit rank-one verification, would considerably aid the reader; Section 7 is a natural place for such a check.","section":"Definition 4.2.4"},{"comment":"The bootstrap in Remark 1.5.4 is easy to misread as circular; a one-sentence dependency diagram or a numbered sequence indicating which results are used before and after Theorem 1.2.2 would make the logical ordering unambiguous.","section":"Remark 1.5.4"},{"comment":"Conjecture 1.7.2 is clearly labelled as a conjecture, but the surrounding discussion in §1.7 could emphasize more strongly that the displayed cycle identity is not a theorem; some results in Section 8 are conditional on this conjecture, and the distinction should be kept explicit.","section":"§1.7 and §8"}],"recommendation":"accept","confidential_remarks":"The main residual risk is the correctness of the new Appendix F characterization; I was not able to verify every line of that appendix during this review. If the editor wishes additional assurance, an independent check of Appendix F, or at least of its rank-one case, would be the highest-value use of further review time. The paper's substantial dependence on the authors' earlier work [EG19] is clearly documented and does not by itself raise novelty concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is the paper that actually resolves the existence of crystalline lifts for every mod p representation of G_K — Theorem 1.2.2 — and it does so by constructing the global moduli stack X_d that Mazur-style deformation rings can only see infinitesimally. Second, the place to look for trouble is Appendix F.\n\nWhat's new: X_d is a Noetherian formal algebraic stack over Spf Z_p, its reduced substack is finite type over F_p, equidimensional of dimension [K:Q_p]d(d-1)/2, with irreducible components labelled by Serre weights. The proof of that equidimensionality is intertwined with the lift theorem—the authors disclose the bootstrap in Remark 1.5.4, and it is a genuine bootstrap, not a circle: the weaker Theorem 5.5.11 suffices for Corollary 1.5.2, and the full strength of Proposition 1.5.1 is recovered afterwards. Theorem 1.2.4 constructs closed crystalline and semistable substacks flat over O, and the geometric Breuil–Mézard statement is honestly labelled a conjecture (1.7.2), with a qualitative version proved in Section 8.\n\nThe architecture is coherent and the paper is unusually candid about its own dependencies. The main load-bearing seam is Definition 4.2.4, the 'admits all descents' criterion for a Breuil–Kisin–Fargues module to be crystalline, proved in Appendix F. Theorem 1.2.4 defines the crystalline substacks through this criterion, and Theorem 1.2.2 relies on effectivity of Spf R^{crys,λ}/p -> X_d, which in turn relies on those substacks being the right ones. If the criterion misclassified—if it were under- or over-inclusive—Corollary 1.5.2 would not control H^2 on the correct locus. I did not find an internal flaw; Appendix F is dense and I did not verify it line by line. The authors themselves remark in Section 1.6 that the compatibilities would be better expressed prismatically, which is consistent with the criterion being subtle rather than wrong. A rank-one explicit check would be a cheap way to settle the worry.\n\nThe paper leans on their earlier [EG19] for Ind-algebraicity of φ-module stacks. That is a legitimate citation, not a flaw, but a referee needs to be comfortable with that dependency.\n\nWho gets value: anyone working on Galois deformation rings, Breuil–Mézard, or Serre weights. It deserves a serious refereeing, and the referees should be pointed to Sections 4, 5, and Appendix F, and to the bootstrap in 1.5.4.","headline":"Emerton–Gee really do construct the global moduli of (φ,Γ)-modules and prove crystalline lifts exist for every mod p representation; the one genuinely load-bearing subtlety is the Appendix F 'admits all descents' criterion.","tokens_in":65415,"tokens_out":3198,"would_cite":true,"duration_ms":33823,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11F80","14D23","11F85"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper constructs global moduli stacks of étale (φ,Γ)-modules and proves from their geometry that every mod p representation of a p-adic Galois group lifts to a crystalline representation with regular Hodge–Tate weights.","keywords":["moduli stacks","étale (φ,Γ)-modules","crystalline lifts","Galois representations","Breuil–Kisin–Fargues modules","Herr complex","Serre weights","deformation rings"],"falsifier":"Take a specific semistable non-crystalline representation of $G_{\\mathbb{Q}_p}$, for instance the $p$-adic representation attached to a split multiplicative Tate curve, and compute its Breuil–Kisin–Fargues module: if for every uniformiser $\\pi$ the module descends to $S_{\\pi^\\flat,\\mathbb{Z}_p}$ and satisfies both independence conditions of Definition 4.2.4, the criterion of Appendix F is false. Alternatively, test Corollary 1.5.2 in a small example such as $\\rho = \\chi_1 \\oplus \\chi_2$ for characters of $G_{\\mathbb{Q}_p}$: compute the locus in $\\mathrm{Spec}\\,R^{\\mathrm{crys},\\lambda}_\\rho/p$ where $\\dim H^2 \\ge r$; if its codimension is ever less than $r$, the main lifting theorem fails.","tokens_in":64030,"feed_emoji":"🧊","tokens_out":15047,"duration_ms":129555,"temperature":0.7,"pith_summary":"The paper constructs global moduli objects—Noetherian formal algebraic stacks $X_d$ over $\\mathrm{Spf}\\,\\mathbb{Z}_p$—whose $A$-valued points are rank $d$ étale $(\\phi,\\Gamma)$-modules with coefficients in a p-adically complete ring $A$. Because étale $(\\phi,\\Gamma)$-modules with finite coefficients are equivalent to continuous representations of $G_K$, the formal completion of $X_d$ at a residual representation $\\rho$ is exactly its universal deformation ring; the stack is the global object that Mazur's local deformation rings sit inside. The paper then uses the geometry of the reduced special fiber—equidimensional of dimension $[K:\\mathbb{Q}_p]d(d-1)/2$ with irreducible components labelled by Serre weights—to prove that every mod $p$ representation $\\rho:G_K\\to\\mathrm{GL}_d(\\mathbb{F}_p)$ admits a lift that is crystalline of regular Hodge–Tate weights and potentially diagonalizable. This is the first proof of such a lifting theorem in full generality, because the local $H^2$ obstructions that prevent lifting extension classes are controlled globally by a codimension bound on the stack.","feed_headline":"Every mod p Galois representation has a crystalline lift","feed_subtitle":"New moduli stacks of (φ,Γ)-modules supply the global geometry that local deformation theory lacked.","key_machinery":"The central object is the formal algebraic stack $X_d$ of étale $(\\phi,\\Gamma)$-modules with coefficients. Three mechanisms carry the argument: (1) the Herr complex, made into a perfect complex of $A$-modules computing Galois cohomology in families, which is used to inductively construct the irreducible closed substacks $X^k_{d,\\mathrm{red}}$ of dimension $[K:\\mathbb{Q}_p]d(d-1)/2$ as families of successive extensions of characters; (2) the crystalline and semistable closed substacks $X^{\\mathrm{crys},\\lambda}_d$ and $X^{\\mathrm{ss},\\lambda}_d$, defined by the criterion that a Breuil–Kisin–Fargues module 'admits all descents' (descends to $S_{\\pi^\\flat,A}$ for every uniformiser $\\pi$, with two independence conditions, proved in Appendix F); (3) the effectivity of the versal morphisms $\\mathrm{Spf}\\,R^{\\mathrm{crys},\\lambda}_\\rho/p\\to X_d$, which holds because $X^{\\mathrm{crys},\\lambda}_d$ is a p-adic formal algebraic stack, so its special fibre is an algebraic stack and versal rings for algebraic stacks are effective. The effectivity converts the $H^2$ codimension bound on $X_{d,\\mathrm{red}}$ into Corollary 1.5.2, a bound on the special fibres of crystalline lifting rings, which is exactly what is needed to lift extension classes to characteristic zero.","core_discovery":"On the paper's own terms: the category fibred in groupoids of rank $d$ projective étale $(\\phi,\\Gamma)$-modules over p-adically complete $\\mathbb{Z}_p$-algebras is a Noetherian formal algebraic stack $X_d$ over $\\mathrm{Spf}\\,\\mathbb{Z}_p$, and its reduced substack $X_{d,\\mathrm{red}}$ is of finite type over $\\mathbb{F}_p$, equidimensional of dimension $[K:\\mathbb{Q}_p]d(d-1)/2$, with irreducible components naturally labelled by Serre weights. The closed substacks $X^{\\mathrm{crys},\\lambda}_d$ and $X^{\\mathrm{ss},\\lambda}_d$ of $(X_d)_O$, defined via a new criterion for crystallinity in terms of Breuil–Kisin–Fargues modules that admit all descents, are p-adic formal algebraic stacks flat over $O$ whose finite-flat $A$-valued points are exactly the representations that become crystalline (resp. semistable) of Hodge type $\\lambda$ after extension of scalars. From the codimension bound of Proposition 1.5.1—the locus of points in $X_{d,\\mathrm{red}}$ with $\\dim H^2 \\ge r$ is Zariski closed of codimension $\\ge r$—together with the effectivity of crystalline versal rings that follows from the p-adic formal algebraic structure, the paper derives the main application (Theorem 1.2.2): every continuous $\\rho:G_K\\to\\mathrm{GL}_d(\\mathbb{F}_p)$ has a lift $\\rho^\\circ:G_K\\to\\mathrm{GL}_d(\\mathbb{Z}_p)$ whose generic fibre is crystalline of regular Hodge–Tate weights, and $\\rho^\\circ$ can be chosen potentially diagonalizable.","pith_inferences":["If the 'admits all descents' criterion is re-cast in the language of prismatic cohomology (the reformulation the paper itself flags in Section 1.6), the crystalline substacks $X^{\\mathrm{crys},\\lambda}_d$ should admit a more canonical description that may behave better under arbitrary base change and for higher-dimensional families; this is an editorial extrapolation, not a claim of the paper.","Proposition 1.5.1 concerns $H^2$ only; the same stack-geometric method should yield analogous codimension bounds for higher derived invariants or for the derived deformation ring, which would give lifting theorems for further properties (e.g., prescribed inertial type) by the same extension-lifting induction.","Because $X_d$ is not a p-adic formal algebraic stack (Proposition 6.5.2), the rigid-analytic generic fibre of $X_d$ cannot be a usual rigid space; the folklore expectation that the lifting rings are complete intersections of dimension $1 + d^2 + [K:\\mathbb{Q}_p]d^2$ suggests a dimension theory for these formal stacks that has not yet been written down.","The compatibility of the Serre-weight labelling of $X_{d,\\mathrm{red}}$ with the geometric Breuil–Mézard cycles $Z_k$ implies that multiplicities in the special fibres of crystalline lifting rings are determined by intersections of $X^{\\mathrm{crys},\\lambda}_d$ with the components of $X_{d,\\mathrm{red}}$; computing these intersections for $d=2$ should recover the known Breuil–Mézard multiplicities"],"forward_implications":["The universal deformation ring $R^\\square_\\rho$ of every mod $p$ representation $\\rho$ is realized as the completion of $X_d$ at the corresponding point; any question about all deformations of $\\rho$ simultaneously is a question about the local geometry of one global stack.","Theorem 1.2.2 supplies the local input for potential automorphy: under $p \\nmid 2d$ every mod $p$ local representation can be matched, at some place dividing $p$, by an automorphic Galois representation of an imaginary CM field (Corollary 6.4.7).","The equidimensional special fibre $X_{d,\\mathrm{red}}$ with Serre-weight-labelled components gives a geometric reformulation of the weight part of Serre's conjecture: the refinement of the labelling by the cycles $Z_k$ of the geometric Breuil–Mézard conjecture predicts that $\\rho$ admits Serre weight $k$ iff the component labelled $k$ contains $\\rho$, a statement proved for $\\mathrm{GL}_2$ by prio","Proposition 1.5.1 and Corollary 1.5.2 give, for each regular Hodge type $\\lambda$, a uniform codimension bound on the locus where $\\dim H^2$ is large in the special fibre of the crystalline lifting ring; this is the missing global input that makes the inductive extension-lifting strategy work in arbitrary rank.","The crystalline and semistable stacks $X^{\\mathrm{crys},\\lambda}_d$ and $X^{\\mathrm{ss},\\lambda}_d$ are p-adic formal algebraic stacks, so their mod $p$ reductions are algebraic stacks; this ties the mod $p$ geometry of crystalline and semistable lifting rings to the irreducible components of $X_{d,\\mathrm{red}}$, providing the geometric Breuil–Mézard theorem (Theorem 8.1.4)."],"supporting_citations":[{"why":"introduced the formal deformation rings of Galois representations that the stacks algebraize.","marker":"[Maz89]"},{"why":"introduced étale (φ,Γ)-modules and the equivalence with Z_p-representations of G_K that defines the stack's points.","marker":"[Fon90]"},{"why":"constructed crystalline deformation rings via Breuil-Kisin modules, providing the rings R^{crys,λ}_ρ around which the effectivity argument is built.","marker":"[Kis08]"},{"why":"gave the explicit complex used to compute Galois cohomology in families and to construct the irreducible components of X_{d,red}.","marker":"[Her98, Her01]"},{"why":"the authors' earlier paper supplies the moduli stacks of φ-modules and the formal-algebraic-stack machinery used to build X_d.","marker":"[EG19]"},{"why":"provides the Frobenius-amplification construction used to canonically extend Galois actions on Breuil-Kisin modules, needed to define the crystalline substacks.","marker":"[CL11]"},{"why":"introduced the A_inf perspective and Breuil-Kisin-Fargues modules that underpin the 'admits all descents' criterion.","marker":"[BMS19]"},{"why":"descent results that verify the stack property for the moduli constructions used throughout.","marker":"[Dri06]"},{"why":"introduced the notion of potentially diagonalizable representations and descent of lifting rings, used in the main lift theorem.","marker":"[BLGGT14]"}],"fun_headline_variants":["Every mod p Galois rep admits a crystalline lift","Crystalline lifts exist for every mod p Galois rep","New moduli stacks show all mod p Galois reps lift crystalline","All mod p Galois reps admit crystalline lifts via stacks","Crystalline lifts for all mod p Galois reps via moduli stacks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the new characterization (Definition 4.2.4, proved in Appendix F) that an integral p-adic representation is crystalline precisely when its Breuil–Kisin–Fargues module admits all descents—descends to $S_{\\pi^\\flat,A}$ for every choice of uniformiser and satisfies two independence conditions; if this criterion failed for some non-crystalline representation, the crystalline substacks would classify the wrong objects and the effectivity step proving the crystalline lift would collapse.","fun_headline_variants_meta":{"raw":{"variants":["Every mod p Galois rep admits a crystalline lift","Crystalline lifts exist for every mod p Galois rep","New moduli stacks show all mod p Galois reps lift crystalline","All mod p Galois reps admit crystalline lifts via stacks","Crystalline lifts for all mod p Galois reps via moduli stacks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00086,"raw_usage":{"total_tokens":3774,"prompt_tokens":1028,"completion_tokens":2746,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":644,"completion_tokens_details":{"reasoning_tokens":2661}},"tokens_in":644,"tokens_out":2746,"duration_ms":20783,"temperature":1.0,"reasoning_tokens":2661,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:24:38.437318+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a specific semistable non-crystalline representation of $G_{\\mathbb{Q}_p}$, for instance the $p$-adic representation attached to a split multiplicative Tate curve, and compute its Breuil–Kisin–Fargues module: if for every uniformiser $\\pi$ the module descends to $S_{\\pi^\\flat,\\mathbb{Z}_p}$ and satisfies both independence conditions of Definition 4.2.4, the criterion of Appendix F is false. Alternatively, test Corollary 1.5.2 in a small example such as $\\rho = \\chi_1 \\oplus \\chi_2$ for characters of $G_{\\mathbb{Q}_p}$: compute the locus in $\\mathrm{Spec}\\,R^{\\mathrm{crys},\\lambda}_\\rho/p$ where $\\dim H^2 \\ge r$; if its codimension is ever less than $r$, the main lifting theorem fails.","supporting_citations":[],"review_version":1}