{"id":"0c517c87-4b4e-4fe8-a908-2b0efa7de2e8","arxiv_id":"2412.08423","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every unit Cartier module over a noetherian F-finite ring of prime characteristic p has injective dimension at most the dimension of its support plus one; the same holds for unit Frobenius modules over regular rings.","lead":"Mathematicians proved a sharp bound on the complexity of certain algebraic objects (unit Cartier and Frobenius modules) in prime characteristic p. Their result removes a technical restriction from an earlier bound and is optimal in several cases, giving a cleaner structural tool for prime characteristic algebra.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.4 rests on Proposition 4.1, whose proof is deferred to an unpublished self-citation and whose printed domain appears to be a typo; as written the derived unit is not established.","rationale":"The reader identified Proposition 4.1 as the weakest assumption, and I agree: the proof of Theorem 4.4 depends on transferring an injective resolution from a regular ring R down to A, and that transfer is exactly the content of Proposition 4.1. My reading adds a concrete observation: as printed, the domain of the unit is inconsistent with the adjunction, since ψ^unit_∗ goes from Cart^unit_A to Cart^unit_R and ψ^♭_unit goes back, so the unit should be on Cart^unit_A, not Cart_R. The surrounding text does not repair this, because Remark 4.2 only identifies Rψ^♭_unit with ψ^! for regular R and does not prove the unit is an isomorphism. The rest of the paper is largely self-contained and internally coherent: Proposition 3.1, Lemma 3.2, Lemma 3.3, and Corollary 3.4 appear to follow from the stated definitions and cited results, and the support argument in Theorem 4.4 is plausible. Thus the conditional verdict is appropriate: the main bound is likely correct, but the manuscript should either provide a proof of the corrected Proposition 4.1 or cite a published, independently verified version before the central claim can be accepted as standalone.","tokens_in":10180,"tokens_out":43203,"duration_ms":448421,"concrete_test":"Take R = F_p[x], A = F_p, and N = F_p with the unique unit Cartier structure. Compute the unit map N → ψ^♭_unit(ψ^unit_∗ N) explicitly using the definitions in §2 and Eq. (3.0.1). If the target is not isomorphic to N, or if uψ_∗N is infinite-dimensional so that the map cannot be an isomorphism, Proposition 4.1 fails. If the map is an isomorphism, the deferred Kashiwara-type step is at least substantiated in the minimal closed-immersion case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central reduction in Lemma 4.3 requires the equivalence N ≃ Rψ^♭_unit(ψ^unit_∗ N) for N ∈ Cart^unit_A. In the printed Proposition 4.1, the unit is written as id on D^+(Cart_R), but the adjunction is ψ^unit_∗ : Cart^unit_A ⇄ Cart^unit_R : ψ^♭_unit, so its unit should be on Cart^unit_A, not Cart_R; Lemma 4.3 applies the statement to N ∈ Cart^unit_A, making the printed version unusable. If the intended statement is the A-sided unit, the proof is only 'Immediate from [BF22, Proposition 9]', and the nearby Remark 4.2 derives a formula for Rψ^♭_unit but does not prove that the derived unit is an isomorphism. Since Theorem 4.4 is exactly the transfer from the regular ring R down to A, this is the load-bearing step. This is a derivability gap rather than a demonstrated error, but it is the precise place where the proof could fail.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper establishes upper bounds for injective dimension in categories of unit Frobenius and unit Cartier modules over F-finite rings of prime characteristic. Theorem A (Theorem 3.5) states that over a regular F-finite ring R, every unit Frobenius module M satisfies inj.dim_{Frob^unit_R}(M) ≤ inj.dim_R(M)+1 ≤ dim(Supp_R M)+1. Theorem B (Theorem 4.4) states that over an arbitrary noetherian F-finite ring A, every unit Cartier module N satisfies inj.dim_{Cart^unit_A}(N) ≤ dim(Supp_A N)+1, and hence the global injective dimension of the category of unit Cartier modules is at most dim A+1. The proof of Theorem A combines a general comparison between Cartier and underlying module injective dimensions (Proposition 3.1), an equality over regular rings (Lemma 3.3), and Lyubeznik's bound. The proof of Theorem B reduces the singular case to the regular case through a Kashiwara-type adjunction induced by a surjection ψ:R→A with R regular (Section 4). The main technical tool is Proposition 4.1, which asserts that a certain derived unit of adjunction is an isomorphism; the proof is deferred to the authors' unpublished preprint [BF22, Proposition 9].","tokens_in":10345,"tokens_out":9518,"duration_ms":101916,"significance":"If the arguments are correct, the results are natural and valuable: Theorem B removes the dependence on a chosen regular presentation in the prior bound of [BF22, Theorem 1], giving an intrinsic bound dim(Supp N)+1, and Remark 4.6 indicates that the bound is optimal for equidimensional Cohen-Macaulay rings. The statements improve on Ma's bound in the regular case and give a clean category-level injective dimension bound over all F-finite rings. The internal arguments in Section 3 are concise and mostly self-contained, and the overall strategy is clear and plausible. However, the central reduction in Section 4 rests on a statement that is both misprinted and not proved within the paper, so the current version is not fully verifiable as written.","major_comments":[{"comment":"The displayed unit is written as η : id_{D^+(Cart_R)} → Rψ^♭_unit ∘ ψ^unit_*. For the adjunction ψ^unit_* : Cart^unit_A ⇄ Cart^unit_R : ψ^♭_unit, the unit should be id_{D^+(Cart^unit_A)} → Rψ^♭_unit ∘ ψ^unit_*; the printed domain is the wrong category and, interpreted literally, the statement is not the unit of this adjunction. Lemma 4.3 applies the proposition to N ∈ Cart^unit_A and needs exactly the A-sided unit, so this is not a cosmetic typo: Theorem 4.4's reduction to the regular ring R depends on this equivalence. In addition, the proof is only 'Immediate from [BF22, Proposition 9]', where [BF22] is an unpublished preprint by two of the authors, and no statement or argument is reproduced. I ask the authors to correct the statement and provide a self-contained proof, or at least a precise statement and proof of the needed result from [BF22] within the paper.","section":"§4, Proposition 4.1"},{"comment":"The lemma is the load-bearing transfer step, but its proof is not complete as written. The proof begins with an injective resolution ψ^unit_* N → I^• of length ≤ k in Cart^unit_R and then uses 'the equivalence N ≃ Rψ^♭_unit ψ^unit_* N of Proposition 4.1'. Because Proposition 4.1 is misstated and its proof is deferred, this equivalence is not established in the manuscript. Moreover, the step from the quasi-isomorphism N ≃ ψ^♭_unit I^• to the conclusion that N has an injective resolution of length ≤ k in Cart^unit_A requires that ψ^♭_unit I^• is a bounded complex of injectives and that the quasi-isomorphism is a genuine resolution of N; the text asserts this in one sentence. The argument is likely standard, but given that this lemma is exactly what makes Theorem 4.4 follow, the details should be spelled out.","section":"§4, Lemma 4.3"},{"comment":"The support equality Supp_A N = Supp_R ψ_* N = Supp_R uψ_* N is used to convert the regular-ring bound into dim(Supp_A N)+1, but the proof is abbreviated. The injectivity of ψ_* N → F^♭ψ_* N is shown, and the equality of supports of ψ_* N and F^♭ψ_* N follows from the localization isomorphism and the support containment; however, the final equality with Supp_R uψ_* N depends on the observation that the colimit u(ψ_* N) contains ψ_* N as a submodule because all transition maps are injective. This is true, but it should be stated explicitly, since it is needed for the dimension estimate.","section":"§4, Theorem 4.4 proof"}],"minor_comments":[{"comment":"In the last display of the proof, '= Supp_A N + 1' should read '= dim Supp_A N + 1'.","section":"§4, Theorem 4.4 proof"},{"comment":"The word 'breifly' is a typo for 'briefly'.","section":"§4, opening sentence"},{"comment":"The remark states that ψ^♭ preserves the unit property and concludes Rψ^♭_unit M = ψ^!M over regular R; this is used only implicitly later, and it would help to state explicitly that ψ^♭ I^• is then a unit Cartier complex, so that uψ^♭ I^• = ψ^♭ I^•.","section":"§4, Remark 4.2"},{"comment":"The phrase 'an noetherian F-finite ring' has an article disagreement; it should be 'a noetherian F-finite ring'.","section":"§1, Theorem 1.3"}],"recommendation":"major_revision","confidential_remarks":"The paper relies heavily on the authors' own unpublished preprint [BF22], not only for background but for the exact transfer statement (Proposition 4.1) on which Theorem B depends. Even if [BF22] is forthcoming, the published version of this paper should either include a full proof of Proposition 4.1 or make the dependence precise and verifiable. The misstatement of the domain of the unit in Proposition 4.1 is another reason the current text cannot be accepted as is; both issues are fixable in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper proves the right bound, and the regular case is solid. The singular case has a hole that is probably patchable but as printed it does not close.\n\nWhat's actually new: Theorem A improves Ma's bound from dim(R)+1 to dim(Supp M)+1 for unit Frobenius modules over regular F-finite R. Theorem B removes the auxiliary regular presentation from the BF22 bound and gives the same support-based bound for unit Cartier modules over any F-finite A. That is a genuine improvement, and the internal machinery – Proposition 3.1, Lemma 3.3, Corollary 3.4 – is clean and self-contained. The equivalence with Cartier crystals is standard, and the support argument in Theorem 4.4 is basically fine modulo a typo in the inclusion direction.\n\nThe soft spot is exactly Proposition 4.1. The statement as printed says the derived unit is an isomorphism on D^+(Cart_R), but the adjunction at issue is between Cart^unit_A and Cart^unit_R, so the unit should be on D^+(Cart^unit_A). Lemma 4.3 uses the A-sided statement. So the printed proposition is unusable as written. The proof is a one-liner citing [BF22, Prop 9], an unpublished arXiv preprint. That is a derivability gap in the load-bearing step, not a demonstrated error; the intended claim is plausible and likely follows from BF22's work. But the manuscript should either prove Proposition 4.1 in the text or cite a published, vetted version.\n\nThere are smaller issues: the support inclusion in the proof of Theorem 4.4 is stated backwards (should be ⊆), and the optimality remark is borrowed from the same preprint. None of these change the main arc.\n\nBottom line: for a specialist in prime-characteristic commutative algebra, this is a useful short paper and it should get a serious referee. I would not desk-reject it. I'd ask the referee to verify the intended form of Proposition 4.1 and insist the authors fix the statement and the proof before acceptance. If the authors can do that, the paper becomes a solid contribution.","headline":"Nice short paper on sharp injective dimension bounds for unit Cartier/Frobenius modules; the regular case is solid, but the singular case rests on a mis-stated, unproven proposition deferred to an unpublished preprint.","tokens_in":10972,"tokens_out":3393,"would_cite":true,"duration_ms":31915,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13D05","13A35","14B15"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that the injective dimension of every unit Cartier module over a noetherian F-finite ring of prime characteristic p is at most the dimension of its support plus one, and similarly for unit Frobenius modules over regular…","keywords":["unit Cartier modules","unit Frobenius modules","injective dimension","F-finite rings","prime characteristic","Cartier crystals","global dimension","Frobenius endomorphism"],"falsifier":"Find a noetherian F-finite ring A and a unit Cartier module N with $\\operatorname{inj.dim}_{\\mathrm{Cart}^{\\mathrm{unit}}_A}(N) > \\dim(\\mathrm{Supp}_A\\, N)+1$. A concrete check would be to compute this category-level injective dimension for a non-Cohen-Macaulay F-finite ring, since the paper's Remark 4.6 already gives equality $\\dim A+1$ for the canonical module over equidimensional Cohen-Macaulay rings.","tokens_in":9929,"feed_emoji":"🧮","tokens_out":7927,"duration_ms":79017,"temperature":0.7,"pith_summary":"The paper establishes a sharp, support-sensitive bound on how many injective steps are needed to resolve modules carrying a Frobenius action. For unit Cartier modules over any noetherian F-finite ring of prime characteristic p, the injective dimension in the category of unit Cartier modules is at most the dimension of the module's support plus one. Over a regular F-finite ring, the analogous statement holds for unit Frobenius modules, with the extra refinement that their injective dimension is at most the ordinary injective dimension of the underlying module plus one. Because the support dimension of a module is often much smaller than the dimension of the ambient ring, this converts an earlier ambient-ring bound into an intrinsic one.","feed_headline":"Unit Cartier modules resolve in support dimension plus one","feed_subtitle":"For any F-finite ring in characteristic p, every unit Cartier module has a bounded injective resolution controlled by its own support.","key_machinery":"The main mechanism is the category of unit Cartier modules as right modules over the twisted polynomial ring $A[F]$, together with the unitalization functor $u(M)=\\operatorname{colim}(M \\to F^\\flat M \\to F^{\\flat 2}M \\to \\cdots)$ that cuts Cartier modules down to unit ones. Over a regular ring, tensoring with the canonical module $\\omega_R$ identifies unit Cartier modules with unit Frobenius modules, so bounds transfer between the two settings. The decisive step for singular A is a Kashiwara-type transfer statement, Proposition 4.1: for a surjection $\\psi:R \\to A$ with R regular F-finite, the derived unit of adjunction $\\mathrm{id} \\to R\\psi^\\flat_{\\mathrm{unit}} \\circ \\psi^{\\mathrm{unit}}_*$ is an isomorphism, so an injective resolution of the pulled-back module over R can be pushed back down to give an injective resolution over A. The bound then follows by comparing supports under pullback and using the regular case.","core_discovery":"The paper's central claim is Theorem B: for every noetherian F-finite ring A of prime characteristic p, every unit Cartier module N satisfies $\\operatorname{inj.dim}_{\\mathrm{Cart}^{\\mathrm{unit}}_A}(N) \\leq \\dim(\\mathrm{Supp}_A\\, N)+1$. Consequently the global injective dimension of the category of unit Cartier modules over A is at most $\\dim A+1$, and this is optimal: for an equidimensional Cohen-Macaulay A, the canonical module $\\omega_A$ has injective dimension exactly $\\dim A+1$ in that category. The paper also proves Theorem A for regular F-finite rings R: every unit Frobenius module M satisfies $\\operatorname{inj.dim}_{\\mathrm{Frob}^{\\mathrm{unit}}_R}(M) \\leq \\operatorname{inj.dim}_R(M)+1 \\leq \\dim(\\mathrm{Supp}_R\\, M)+1$. The first inequality sharpens the known bound $\\operatorname{inj.dim}_R(M) \\leq \\dim(\\mathrm{Supp}_R\\, M)$ for the underlying module, and the second shows that the bound depends only on the support.","pith_inferences":["Beyond the paper: if the bound is correct, injective dimension in the unit-Cartier category is essentially a dimension-of-support invariant, so this number alone cannot separate unit Cartier modules whose supports have the same dimension.","A natural extension not tested here would be to drop the F-finiteness hypothesis; the proof uses a structure result that presents F-finite rings as quotients of regular rings, and the same support-plus-one bound is not established without F-finiteness.","The transfer statement quoted from the earlier preprint could be tested directly on a concrete nonregular quotient, such as $A = k[x,y]/(xy)$, by checking whether the derived unit is a quasi-isomorphism on a known unit Cartier module; this would exercise the load-bearing step on a small example."],"forward_implications":["The category of unit Cartier modules over any noetherian F-finite ring A has global injective dimension exactly $\\dim A+1$ whenever A is equidimensional Cohen-Macaulay, and at most $\\dim A+1$ in general.","For unit Frobenius modules over a regular F-finite ring, the injective dimension in their own category is bounded by $\\operatorname{inj.dim}_R(M)+1$, which is strictly sharper than $\\dim(\\mathrm{Supp}_R\\, M)+1$ whenever the underlying module is not already extremal.","Because unit Cartier modules form an abelian category for singular F-finite rings, the bound gives a finite, dimension-controlled resolution theory for modules arising from the Frobenius action, including injective hulls of residue fields.","The same bound holds for quasi-coherent Cartier crystals, since that category is equivalent to unit Cartier modules, as noted in Remark 4.5.","The result removes dependence on a chosen regular presentation: earlier bounds depended on a surjection $R \\to A$, while the new bound is intrinsic to A."],"supporting_citations":[{"why":"Supplies the foundational theorem that a unit Frobenius module over a regular ring has ordinary injective dimension bounded by the dimension of its support, which Theorem A sharpens.","marker":"[Lyu97]"},{"why":"Provides the equivalence between unit Frobenius modules and unit Cartier modules over regular F-finite rings and the earlier finite-global-dimension bound that the paper refines.","marker":"[Ma14]"},{"why":"Supplies the unitalization functor, the Cartier-crystal equivalence, the earlier ambient-ring bound for unit Cartier modules, and the transfer statement used as Proposition 4.1.","marker":"[BF22]"},{"why":"Records that every F-finite noetherian ring is a homomorphic image of a regular F-finite ring, used to choose the surjection in the proof of Theorem B.","marker":"[Gab04]"},{"why":"Establishes that $F_*R$ is projective over a regular ring R, which makes $F^\\flat$ exact and makes the inclusion of unit Cartier modules into all Cartier modules exact.","marker":"[Kun69]"},{"why":"Defines the Cartier-module formalism, the equivalent presentations of structure maps, and the category-level background used throughout Sections 2 and 4.","marker":"[BB13]"}],"fun_headline_variants":["Injective dimension bounded by support size plus one","Unit Cartier and Frobenius modules: injective dim ≤ support + 1","Support dimension + 1 caps injective dimension","Sharp bound: injective dimension is support + 1 for Cartier modules","Injective dimension ≤ support dimension + 1 for unit modules"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof relies on a transfer statement, quoted from an earlier preprint and not proved in this text: pulling a module up from a quotient ring to a regular ring and pushing it back down recovers it up to quasi-isomorphism. If that statement failed, the main inequality for singular rings would not follow from the argument given.","fun_headline_variants_meta":{"raw":{"variants":["Injective dimension bounded by support size plus one","Unit Cartier and Frobenius modules: injective dim ≤ support + 1","Support dimension + 1 caps injective dimension","Sharp bound: injective dimension is support + 1 for Cartier modules","Injective dimension ≤ support dimension + 1 for unit modules"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000339,"raw_usage":{"total_tokens":1845,"prompt_tokens":894,"completion_tokens":951,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":510,"completion_tokens_details":{"reasoning_tokens":863}},"tokens_in":510,"tokens_out":951,"duration_ms":9719,"temperature":1.0,"reasoning_tokens":863,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T17:51:59.311129+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a noetherian F-finite ring A and a unit Cartier module N with $\\operatorname{inj.dim}_{\\mathrm{Cart}^{\\mathrm{unit}}_A}(N) > \\dim(\\mathrm{Supp}_A\\, N)+1$. A concrete check would be to compute this category-level injective dimension for a non-Cohen-Macaulay F-finite ring, since the paper's Remark 4.6 already gives equality $\\dim A+1$ for the canonical module over equidimensional Cohen-Macaulay rings.","supporting_citations":[],"review_version":1}