{"id":"595cdf31-0b24-4072-84b8-6b0371b6178e","arxiv_id":"1908.08262","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Arapura gives a simpler construction of motivic sheaves and a simpler proof that the Leray spectral sequence has a motivic lift whose Betti realization is the classical Leray spectral sequence.","lead":"An algebraic geometer simplifies the construction of Nori-based motivic sheaves and proves that the Leray spectral sequence of a projective map can be lifted into this universal category. The result makes the motivic cohomology machine more transparent and gives a new motivic decomposition statement for smooth projective maps.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central construction rests entirely on Lemma 5.2, quoted without proof from [A1]; if that lemma fails over arbitrary noetherian R, the cellular method and Theorem 6.1 collapse.","rationale":"The reader's conditional verdict identifies Lemma 5.2 as the weakest assumption; I agree. I did not find an independent fatal flaw: the overall strategy is coherent, the N+ formalism is applied consistently, and the Betti-realization checks are mostly sound given an exact faithful Betti functor. The main theorem's proof does contain a local sign/direction issue in the type-II compatibility diagram, where the displayed morphism in Proposition 6.3 is reversed relative to (5.5), but this appears to be a correctable typo rather than a conceptual gap. The truly load-bearing point remains the unproved cellular lemma: the paper provides no self-contained proof and explicitly delegates it to [A1]. Because the current paper works over arbitrary noetherian rings and all constructible sheaves, verifying that the quoted lemma carries this generality and that its refinement clause supports the directed-union argument is essential. Since this matches the reader's concern and I have no additional objection, the verdict should stay unchanged.","tokens_in":15388,"tokens_out":40019,"duration_ms":401158,"concrete_test":"Independently re-derive [A1, Lemma 3.7] from its stated hypotheses and check the coefficient ring: the proof must work for constructible sheaves of R-modules for arbitrary commutative noetherian R, not just Q-vector spaces, and the 'refine a given chain' clause must hold for finitely many sheaves simultaneously. A concrete computation to run is T = A^2, R = Z, F the constant sheaf, with a prescribed chain whose T_1 contains a smooth cubic; determine whether an enlargement of T_1 exists with H^1(T_2, T_1; F) = 0 and H^2(T_1, T_0; F) = 0. This tests both the vanishing and the refinement claim that underpin Corollary 5.3 and Proposition 6.3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 5.2 is the unproved load-bearing input. It asserts that for every constructible sheaf F on the affine Jouanolou cover T there is a chain T_{-1} ⊂ T_0 ⊂ ... ⊂ T of equidimensional closed sets (dim T_a = a) with H^i(T_a, T_{a-1}; F) = 0 unless i = a. This vanishing is what makes Lemma 5.4 true (via the de Cataldo-Migliorini criterion), what identifies the complex K(i)^• with H^*(S, H^i_S(X, Y)) in Corollary 5.5, and what makes the complexes K_{T•}(X, Y, i) in Proposition 6.3 compute the correct graded pieces. The lemma is quoted as [A1, Lemma 3.7] with no proof, and the abstract states that the paper relies on prior work for some technical results. The current paper fixes an arbitrary commutative noetherian ring R and works with all k-constructible sheaves; a referee must confirm that [A1] really proves the lemma in this generality, in particular for R = Z and not only Q-vector spaces, and that the 'refine a given chain' clause can be iterated to yield the directed-union Corollary 5.3. If either fails, the cellular method defining h^* has no replacement in this paper. A smaller, repairable issue is that the type-II diagram in Proposition 6.3 appears to have its arrows reversed relative to (5.5); this should be corrected when checking the representation property.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a simplified construction of the author's category of motivic sheaves over a k-variety S. It starts from the N+ construction of Barbieri Viale and Prest, applied to a quiver Δ(S) whose vertices are triples (X→S, Y, i) satisfying a base change property, with Betti representation H^i_S(X,Y;R). Inverting the Lefschetz motive yields the category M(S,R). The main theorem (Theorem 6.1) asserts a δ-functor h^*:M(S)→M(k) compatible with Betti realization, and, for controlled pairs, a motivic spectral sequence whose Betti realization is the Leray spectral sequence. The paper also sketches a Hodge realization over a curve and derives a noncanonical decomposition for smooth projective maps and for maps over a smooth projective curve.","tokens_in":15693,"tokens_out":8509,"duration_ms":83914,"significance":"If the main theorem is correct, it provides a canonical motivic lift of the Leray spectral sequence and simplifies the earlier constructions of [A1] and [A2], with coefficients in an arbitrary commutative noetherian ring. The use of the N+ construction and of the de Cataldo–Migliorini criterion is a genuine simplification, and the compatibility with the Betti realization is checked through faithful exact functors. The proof is detailed and the main construction is explicit. However, the central construction leans on Lemma 5.2, quoted from earlier work, and the manuscript does not demonstrate that this lemma holds in the full generality needed here; this is the main risk to the paper's central claim.","major_comments":[{"comment":"Lemma 5.2 is the central load-bearing input of the paper: it is used to prove Lemma 5.4 via the de Cataldo–Migliorini criterion, to identify K(i)^• with H^*(S,H^i_S(X,Y)) in Corollary 5.5, and to define the complexes K_{T•}(X,Y,i) in Proposition 6.3 and the spectral sequence in Theorem 6.1. The proof given is only a citation to [A1, Lemma 3.7]. Since the paper works with an arbitrary commutative noetherian ring R and with all k-constructible sheaves, the authors should either include a proof or state precisely the hypotheses under which [A1] proves the lemma, and confirm that [A1] covers this coefficient generality and the 'refine a given chain' clause needed for Corollary 5.3. Without this, the construction of h^* has no self-contained replacement.","section":"§5, Lemma 5.2"},{"comment":"The theorem is stated for controlled pairs (f:X→S,Y), but the proof constructs the exact couple from complexes h^*(X_{T_p},Y_{T_p}∪X_{T_{p-1}}) without explicitly proving that a controlled pair is cellular with respect to the chosen chain T•. Corollary 5.3 only places the object in some Δ(S,T•); one still needs the vanishing condition (5.1) for the particular pair in order to identify the motivic spectral sequence with the Leray spectral sequence. Please add the argument that, for a controlled pair, Lemma 5.2 provides a chain with respect to which the pair is cellular, or restrict the statement and then explain the passage.","section":"§6, proof of Theorem 6.1"},{"comment":"Lemma 5.2 is stated for a constructible sheaf F on T, but the cellularity condition (5.1) and the object F=π^*H_S(X,Y;R) in Lemma 5.4 involve a complex of sheaves. The paper does not spell out how the sheaf-level vanishing of Lemma 5.2 is converted into the hypercohomology vanishing needed for Lemma 5.4 and Corollary 5.5. This conversion is a nontrivial step and should be made explicit.","section":"§5, Eq. (5.1) and Lemma 5.2"}],"minor_comments":[{"comment":"The arrows in the diagram for a type-II morphism appear to be reversed relative to (5.5): the map to be constructed is K_{T•}(Y,Z,i)→K_{T•}(X,Y,i+1), whereas the display, as written, suggests the opposite direction. Please correct the diagram and the accompanying indexing.","section":"§6, Proposition 6.3"},{"comment":"The name 'Jounalou' should be 'Jouanolou' in the statement of Lemma 5.1 and in the reference [J].","section":"§5, Lemma 5.1 and references"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern raised by the reader is legitimate: the proof of Theorem 6.1 is anchored by Lemma 5.2 from [A1], and no proof or precise general statement appears here. I would ask the authors to supply the missing verification before publication. There is no indication of circularity; the reliance is external, not on the theorem being proved. The paper fits the journal's scope, and the simplification it offers is valuable if the cited lemma indeed holds in the stated generality."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper does what it says: it rebuilds Arapura's motivic sheaves using the Barbieri-Viale–Prest N+ construction, gives a shorter proof of the motivic Leray spectral sequence, and adds a motivic decomposition theorem for smooth projective maps and curves. That is real progress. The new category M(S) is cleaner than the earlier single-step construction, and the use of the de Cataldo–Migliorini criterion simplifies the homological algebra substantially. I read the proof of Theorem 6.1 as coherent; the exact couple is set up correctly and the Betti-realization comparison is convincing.\n\nThe main soft spot is not hidden: it is Lemma 5.2, quoted from [A1, Lemma 3.7]. The entire cellular method—Corollary 5.5, the complexes K(i)•, and hence the h^j functors—depends on the existence of a chain with the stated vanishing for arbitrary constructible sheaves over a noetherian ring R. The paper says it relies on prior work for technical results, so this is a disclosed dependency, but a referee should check that [A1] proves the lemma in exactly this generality (especially R = Z, not just Q-vector spaces) and that the \"refine a given chain\" clause can be iterated to get the directed union in Corollary 5.3. If that lemma is fine, the main theorem stands; if not, there is no replacement in this paper. That makes the verdict conditional, not negative.\n\nThere is also a small sign/direction issue: in Proposition 6.3 the type-II diagram appears to have the arrows reversed relative to (5.5). I think it is a typo—the intended map should be K_T•(Y,Z,i) → K_T•(X,Y,i+1)—but it should be fixed when the proposition is checked.\n\nThe citation pattern is normal; Arapura cites his own earlier work because he is explicitly reworking it, and the key external inputs [BP], [CM], and [D2] are standard. Otherwise, I found no internal contradiction. The decomposition theorem 6.5 is a straightforward consequence of semisimplicity of pure motives plus the motivic spectral sequence, and it is a nice bonus.\n\nThis is a paper for people who work with Nori motives and motivic sheaves. It deserves a serious referee. It is not a desk reject, and it should be sent out with the explicit instruction to verify Lemma 5.2 in [A1] and to clean up the diagram in Proposition 6.3.","headline":"A genuine simplification of motivic sheaves with a clean motivic Leray spectral sequence, but the referee must verify the quoted cellular lemma before the main construction is fully load-bearing.","tokens_in":16235,"tokens_out":4285,"would_cite":true,"duration_ms":40435,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F42","14F05","14C30"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every quasiprojective variety $S$ has a $\\delta$-functor $h^*$ from motivic sheaves on $S$ to mixed motives whose Betti realization is cohomology of the base, and that this functor carries a motivic version of the…","keywords":["motivic sheaves","Leray spectral sequence","mixed motives","Betti realization","delta-functor","constructible sheaves","cellular decomposition","Hodge realization"],"falsifier":"A concrete way to test the claim is to take a constructible sheaf $F$ on an affine bundle $T$ and attempt to refine a prescribed chain to a cellular chain with the vanishing $H^i(T_a,T_{a-1};F)=0$ for $i\\neq a$; the first example where no such refinement exists falsifies Lemma 5.2 and therefore the construction of $h^j$. Alternatively, for a smooth projective family $f:X\\to\\mathbb{P}^1$, compute the motivic page $h^p(h^q_S(X))$ by the paper's prescription and check that its Betti realization equals $H^p(\\mathbb{P}^1,R^qf_*\\mathbb{Z})$ at every $(p,q)$.","tokens_in":15168,"feed_emoji":"🧩","tokens_out":17545,"duration_ms":149158,"temperature":0.7,"pith_summary":"The paper proves that the Leray spectral sequence of a projective morphism is not merely a computational tool in sheaf cohomology: it has a genuine shadow inside the abelian category of mixed motives. The main result constructs, for every quasiprojective $k$-variety $S$ (with $k\\subset\\mathbb{C}$), a $\\delta$-functor $\\{h^j:\\mathcal{M}(S)\\to\\mathcal{M}(k)\\}$ whose Betti realization is ordinary cohomology $H^j(S,-)$ of the base. For a controlled pair $(f:X\\to S,Y)$, this gives a spectral sequence $M E_2^{pq}=h^p(h^q_S(X,Y))\\Rightarrow h^{p+q}(X,Y)$ in $\\mathcal{M}(k)$ whose image under the Betti realization is the classical Leray spectral sequence. A derived-category version yields a functor $r\\Gamma:D^b\\mathcal{M}(S)\\to D^b\\mathcal{M}(k)$, and in the smooth projective case (or over a curve) the motivic Leray sequence degenerates and splits the motive of $X$ into a sum of fiber motives.","feed_headline":"Motivic Leray spectral sequence realized in mixed motives","feed_subtitle":"A new functor h* turns fiber cohomology into genuine motives; the classical Leray spectral sequence is its shadow.","key_machinery":"The load-bearing object is the quiver (directed graph) $\\Delta(S)$ whose vertices are triples $(X\\to S,Y,i)$ standing for the relative cohomology symbol $h^i_S(X,Y)$, with type-I edges given by geometric morphisms and type-II edges by connecting maps coming from the exact sequence for a triple $Z\\subset Y\\subset X$. Feeding the Betti representation $H:\\Delta(S)^{\\mathrm{op}}\\to \\mathrm{Cons}(S^{\\mathrm{an}},R)$ into the N+ construction—the universal way to convert a quiver representation into an abelian category with an exact faithful functor—produces the effective motivic sheaves $\\mathcal{M}^{\\mathrm{eff}}(S,R)$; inverting the Lefschetz motive $L=h^1_S(\\mathbb{G}_{m,S},1)$ gives $\\mathcal{M}(S,R)$. The proof of the main theorem rests on a cellular decomposition: an affine-bundle reduction replaces $S$ by an affine $T$, and a quoted lemma supplies a chain $T_{-1}\\subset T_0\\subset\\cdots\\subset T$ of equidimensional closed sets with $H^i(T_a,T_{a-1};F)=0$ unless $i=a$. With that vanishing, a filtered-derived-category criterion identifies the truncation filtration with the d\\'ecalage of the support filtration, producing the complexes $K(i)^\\bullet$ and the isomorphism $H^j(K(i)^\\bullet)\\cong H^j(S,h^i_S(X,Y))$; the same exact couple in $\\mathcal{M}(k)$ gives the motivic Leray spectral sequence.","core_discovery":"On its own terms, the paper claims that the category $\\mathcal{M}(S)$ of motivic constructible sheaves over $S$ can be built in one step from the universal exact-faithful construction applied to a quiver of relative cohomology symbols $h^i_S(X,Y)$, and that this category admits a natural $\\delta$-functor $h^*$ to the category of mixed motives over $k$ satisfying $R_B(h^j(M))\\cong H^j(S,R_B(M))$ for every object $M$. The proof fixes an affine bundle $T\\to S$ and a cellular chain $T_\\bullet$, constructs complexes $K(i)^\\bullet$ of motives whose Betti realizations are the complexes appearing in the Leray filtration, and proves a quasi-isomorphism $H^j(K(i)^\\bullet)\\cong H^j(S,h^i_S(X,Y))$. The same exact-couple construction, repeated inside the motivic category, produces the motivic Leray spectral sequence. The paper also derives a triangulated lift $r\\Gamma$ and, in the smooth projective and curve cases, a noncanonical motivic decomposition $h^i(X)\\cong\\bigoplus_{p+q=i}h^q(h^p_S(X))$.","pith_inferences":["The proof's dependence on the affine-bundle reduction and on the quoted cellular vanishing suggests that quasiprojectivity of $S$ is essential; the existence of $h^*$ over non-quasiprojective bases is not settled here and may be the natural next obstruction.","The same exact-couple method can plausibly be iterated to define motivic direct images $f_*$ for arbitrary morphisms $f:X\\to S$, not only maps to a point; the paper only performs the base-field case, so a motivic six-functor formalism remains to be built on this foundation.","Since the Betti realization is faithful but not full, several motivic spectral sequences could in principle share the same Leray shadow; the paper selects one that is compatible with refinements of the cellular chain, which raises a uniqueness question it does not address.","A testable extension is to ask whether the noncanonical splitting in the degeneration theorem can be made functorial in the family, for instance by attaching a motivic vanishing-cycle or monodromy invariant that controls the choice of splitting."],"forward_implications":["For any controlled pair, the terms $h^p(h^q_S(X,Y))$ of the motivic Leray spectral sequence are genuine mixed motives, not just placeholders for cohomology groups; the classical Leray sequence is their Betti realization.","Corollary 6.4 upgrades the $\\delta$-functor to a triangulated functor $r\\Gamma:D^b\\mathcal{M}(S)\\to D^b\\mathcal{M}(k)$, so the whole derived category of motivic sheaves has a well-defined motivic cohomology of the base, with the $h^j$ as its cohomology functors.","When $f:X\\to S$ is smooth projective, or when $S$ is a smooth projective curve, the motivic Leray spectral sequence degenerates at $E_2$ and gives a noncanonical decomposition $h^i(X)\\cong\\bigoplus_{p+q=i}h^q(h^p_S(X))$ in $\\mathcal{M}(k,\\mathbb{Q})$.","The construction is compatible with refinements of the cellular chain, so the resulting functors $h^j$ do not depend on the auxiliary affine bundle and chain chosen in the proof.","The category is built over any commutative noetherian coefficient ring $R$, so the motivic Leray statement holds integrally and not only over a field."],"supporting_citations":[{"why":"Earlier proof that the Leray spectral sequence is motivic; the present paper quotes its Lemma 5.2 on cellular vanishing as the key technical input.","marker":"[A1]"},{"why":"Previous construction of the category of motivic sheaves whose main theorems and realization functors are simplified and reused here.","marker":"[A2]"},{"why":"Supplies the N+ construction, the universal abelian-category construction from a quiver representation that builds the effective motivic sheaves.","marker":"[BP]"},{"why":"Provides the detailed background account of the category of mixed motives used in Section 4.","marker":"[HM]"},{"why":"Gives the filtered-derived-category isomorphism criterion that underlies Lemma 5.4 and the comparison of spectral sequences.","marker":"[CM]"},{"why":"Provides the affine-bundle reduction that lets cohomology be computed on an affine variety over the base.","marker":"[J]"},{"why":"Supplies the d\\'ecalage of filtrations and the standard identification of the Leray spectral sequence with the filtration spectral sequence.","marker":"[D2]"},{"why":"Gives the exact-couple construction of spectral sequences used to build both the topological and the motivic spectral sequences.","marker":"[W]"}],"fun_headline_variants":["Motivic sheaves get a simpler definition and direct image proof","One-step motivic sheaves: a new category from relative cohomology","Exact couples realize the motivic Leray spectral sequence","Simpler motivic sheaves: new proof of direct images","Leray sequence becomes genuinely motivic via exact couples"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction hinges on a quoted lemma from the earlier paper: every reasonably behaved sheaf of coefficients on the affine replacement $T$ admits a nested chain of algebraic closed subsets $T_{-1}\\subset T_0\\subset\\cdots\\subset T$ such that the cohomology of $T_a$ relative to $T_{a-1}$ vanishes except in degree $a$. The paper does not prove this lemma, and if this vanishing failed, the complexes that define $h^j$ would no longer compute Leray cohomology, so the whole construction would collapse.","fun_headline_variants_meta":{"raw":{"variants":["Motivic sheaves get a simpler definition and direct image proof","One-step motivic sheaves: a new category from relative cohomology","Exact couples realize the motivic Leray spectral sequence","Simpler motivic sheaves: new proof of direct images","Leray sequence becomes genuinely motivic via exact couples"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000588,"raw_usage":{"total_tokens":2730,"prompt_tokens":885,"completion_tokens":1845,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":501,"completion_tokens_details":{"reasoning_tokens":1757}},"tokens_in":501,"tokens_out":1845,"duration_ms":14048,"temperature":1.0,"reasoning_tokens":1757,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:44:40.806870+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete way to test the claim is to take a constructible sheaf $F$ on an affine bundle $T$ and attempt to refine a prescribed chain to a cellular chain with the vanishing $H^i(T_a,T_{a-1};F)=0$ for $i\\neq a$; the first example where no such refinement exists falsifies Lemma 5.2 and therefore the construction of $h^j$. Alternatively, for a smooth projective family $f:X\\to\\mathbb{P}^1$, compute the motivic page $h^p(h^q_S(X))$ by the paper's prescription and check that its Betti realization equals $H^p(\\mathbb{P}^1,R^qf_*\\mathbb{Z})$ at every $(p,q)$.","supporting_citations":[],"review_version":1}