REVIEW 3 major objections 4 minor 5 references
Lefschetz morphisms on singular cohomology and local cohomological dimension of toric varieties
T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read This paper shows that the cup product with the first Chern class on the singular cohomology of a toric variety is fully described by fan combinatorics, and uses that description to prove that the local cohomological defect is not a…
desk verdict A genuinely useful paper on toric local cohomological defect whose headline theorem is under-proved: the bridge from the fan computation to Theorem 1.4 needs to be written out. 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 Ishida complex $\mathrm{Ish}^l_\sigma$, a finite-dimensional complex of vector spaces built purely from the faces of a cone, is the central object; it is the degree-zero part of the Grothendieck dual of the Du Bois complex of reflexive differentials. The argument identifies the Grothendieck dual of the Chern class map with a connecting homomorphism in a short exact sequence of Ishida-type complexes, so the cup product becomes a purely combinatorial linear-algebra operation. Theorem 1.4 is the resulting long exact sequence linking the cone's Ishida cohomology to the Lefschetz morphism $c_1$ on the projective exceptional divisor.
What would settle it
Compute $H^2(\mathrm{Ish}^3_\sigma)$ for the two 14-ray cones in Example 1.7; if the two values turn out to be equal, the claim that the local cohomological defect is non-combinatorial is false. The same computation also tests the four-dimensional criterion, since lcdef $=1$ is equivalent to $\dim H^2(E,\mathbb{C}) \geq 2$.
Extended reading notes
Core claim
The central claim is Theorem 1.4: for an $n$-dimensional affine toric variety $X$ attached to a full-dimensional cone $\sigma$, inserting an interior rational ray $\rho$ gives a toric morphism $\pi: \widetilde{X} \to X$ whose exceptional fibre $E$ is a projective toric variety of dimension $n-1$, and the cohomology of the Ishida complex $\mathrm{Ish}^l_\sigma$ sits in a long exact sequence with $H^*(E, \mathrm{Ish}^{*}_E)$, whose maps are dual to the Chern class map of $(-E)|_E$. This makes the cohomological defect of the cone readable from the Lefschetz behaviour of a single projective toric variety. From this the authors derive a four-dimensional criterion (lcdef $= 1$ iff $\dim H^2(E,\mathbb{C}) \geq 2$), and exhibit two cones with identical combinatorial data but different defects, proving non-invariance.
Load-bearing premise
The argument leans on two statements imported from the authors' preceding preprint, not reproved here: the cohomology of any proper toric variety is of a very restricted Hodge type, and the final differential of the Ishida complex is surjective in high degrees; the paper's injectivity, surjectivity, and non-invariance conclusions depend on them.
Editorial extensions
If this is right
- The local cohomological defect of an affine toric variety can be computed from the maximal defect over faces of its cone, so the problem becomes algorithmic.
- The defect is not a combinatorial invariant: Example 1.7 gives two cones with the same fan combinatorics but defects $1$ and $0$.
- Taking a pyramid over a cone leaves the defect unchanged, so one can build $n$-dimensional affine toric varieties with any defect between $0$ and $n-3$.
- In dimension four the defect is $1$ exactly when the exceptional projective threefold has $\dim H^2(E,\mathbb{C}) \geq 2$; otherwise it is $0$.
- The last differential of the Ishida complex vanishes in high degrees, recovering a prior vanishing result via a Hard Lefschetz-type injectivity statement.
Reading between the lines
- Because lcdef is computed from finite-dimensional vector spaces attached to a cone, the same long exact sequence could support an algorithm that decides the defect without resolving singularities; the computer algebra checks in the paper are a first step in that direction.
- The fan-theoretic description of $c_1$ may extend to other natural classes on toric varieties, such as products of Chern classes or other cohomology operations, by iterating the same extension-of-complexes construction.
- The non-invariance example suggests that any purported combinatorial formula for lcdef must use the actual positions of rays, not just the face poset; classifying the extra geometric data needed would be a natural next step.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes an explicit fan-theoretic description of the cup product with the first Chern class of a line bundle on a proper toric variety, via the Grothendieck dual of the Atiyah class and Ishida complexes. The main result, Theorem 1.4, states that for an affine toric variety X associated to a cone sigma, inserting an interior ray rho yields a projective toric variety E such that the cohomology of the Ishida complex of sigma is governed by a long exact sequence involving the cohomology of Ishida complexes of E and the map dual to c_1((-E)|_E). From this the authors derive a criterion for the local cohomological defect in dimension four, prove that lcdef is not a combinatorial invariant, prove invariance under pyramids, and give several combinatorial criteria for lcdef in dimension four.
Significance. If the main theorem is correct, it is a substantial result: it gives a genuinely combinatorial description of the Lefschetz morphism on singular cohomology of toric varieties and connects it to the local cohomological defect of affine toric varieties. The paper also contains an attractive set of applications, including the non-invariance of lcdef and a pyramid-invariance theorem. The algebraic development in Sections 3.3 and 3.4 is detailed, and the exact sequences there are concrete and checkable. The main weakness is that the proof of the headline theorem is not actually written down: the bridge from the projective construction to the affine cone over (E, (-E)|_E) is asserted in one sentence, and Proposition 3.7 is likewise asserted rather than proved. Several load-bearing statements are imported from the authors' unpublished preprint [KV25], which further limits the self-containedness of the paper.
major comments (3)
- [§3.5, Proposition 3.7] Proposition 3.7 is not proved. The entire proof is the sentence that Ish^{p+1}_sigma is the middle term in the short exact sequence of Corollary 3.6. That identification is not shown. The complex of Corollary 3.6 has terms indexed by both e_tau and b_tau, whereas the cone sigma, as stated in §3.5, has only the b_tau faces. One needs an explicit quasi-isomorphism, or at least a coherent argument identifying Ish^{p+1}_sigma with the middle term of that exact sequence. Without this identification, the equivalence between the vanishing of H^l(Ish^{p+1}_sigma) and the stated injectivity/surjectivity conditions on c_1(D) is unsupported.
- [§3.5, proof of Theorem 1.4] The proof of Theorem 1.4 is reduced to a single sentence: "rephrasing Corollaries 3.5 and 3.6 immediately gives a proof of Theorem 1.4." This is not a formality. One must prove that, after placing the interior ray rho on a coordinate axis, the star of rho in the subdivided fan modulo <rho> is the fan of E; that the original cone sigma is identified with the cone over (E, (-E)|_E); that the support function defining sigma is that of the restriction of -E to E; and that Ish^l_sigma is exactly the middle term of Corollary 3.6, with the connecting homomorphism equal to the dual of c_1((-E)|_E). The last point fixes the sign in the long exact sequence, and a sign error would change all lcdef computations in Corollary 1.6 and Example 1.7. These steps need to be written out.
- [§1, Proposition 1.5, Corollary 1.6, and paragraph after Theorem 1.8] Several results that are load-bearing for the paper's applications are taken from the authors' previous preprint [KV25] without proof. These include the statement that singular cohomology of a proper toric variety is mixed of Hodge-Tate type, the surjectivity of the last differential of Ish^p_X for p >= n/2, and the equality lcdef = m for (m+3)-dimensional non-simplicial toric varieties with isolated non-simplicial locus. Since [KV25] is an unpublished preprint, the current manuscript's main applications depend on results that are not verifiable from the present text. The authors should either include precise proofs of these facts in an appendix or replace the dependence with a published reference.
minor comments (4)
- [Examples 1.7 and 4.4] The claims that the two cones in Example 1.7 can be checked with Macaulay2, and that Example 4.4 has dim H^1 = dim H^2 = 1, would be much more useful if the actual Macaulay2 script or the output were included; as written, these computations are an assertion about a non-combinatorial invariant and should be reproducible.
- [§2.5 and Theorem 1.4] The notation Ish^l_E in Theorem 1.4 and Corollary 1.6 is used for the cohomology of the Ishida complex of a projective toric variety E. Since the earlier definition of Ish^l_X is for the sheaf complex, it would help to explicitly state that H^i(E, Ish^l_E) denotes the hypercohomology of the sheaf complex, not the cohomology of the finite-dimensional complex Ish^l_P.
- [§4, Lemma 4.2] The proof of Lemma 4.2 is dismissed as "analogous" to Lemma 4.1. Since this lemma is used in the proof of Theorem 1.8, a short proof or a precise reference to the analogous argument should be included.
- [§3.2, displayed diagram] The commutative diagram in §3.2 is hard to read because the arrows are not labeled and the vertical maps are not named. Labeling the maps would make the comparison with the exact sequence of complexes in §3.3 considerably clearer.
Circularity Check
No circular reduction found; main derivation is self-contained, with secondary reliance on the authors' earlier preprint [KV25] that is not itself circular.
full rationale
The central derivation is not circular. Theorem 1.4 is obtained by reversing the cone construction of Section 3.5 and applying Corollaries 3.5-3.6, which are built from the Atiyah class and Ishida-complex duality against external sources [Ish87], [MP22], [CLS11]. The proof of Theorem 1.4 is terse: the text says 'rephrasing Corollaries 3.5 and 3.6 immediately gives a proof of Theorem 1.4' and omits the explicit identification of the affine cone with the cone over (E, (-E)|_E), including the sign and divisor class. This is a proof gap (a correctness risk), not a circular reduction: the identification is a nontrivial mathematical claim, not a definitional equality that makes the conclusion equal to the input. The paper does rely on several results from the authors' own preprint [KV25] -- H^3(Ish^3_sigma)=0 in Corollary 1.6, the Hodge-Tate assertion in Proposition 1.5, and the starting lcdef=m examples after Theorem 1.8. These are load-bearing for those secondary statements, but they are separate theorems from the present paper's claims and the non-invariance example is also verified by Macaulay2. No parameter is fitted and no 'prediction' is renamed from an input. The main new results (Theorem 1.4, Theorem 1.8, and the combinatorial corollaries) have independent content. Hence no substantive circularity; the score reflects the unverified self-citation imports without treating them as circular.
Assumptions & free parameters
assumptions (7)
- domain assumption Mustata-Popa criterion: lcdef(X) is computed from Ext groups of the Du Bois complex (Theorem 1.1, [MP22, Cor. 5.3]).
- domain assumption Ishida complex duality: RHom_{O_X}(Omega^p_X, omega_X) is isomorphic to Ish^{n-p}_X for toric varieties ([Ish87], Prop. 2.4).
- domain assumption Du Bois complex of a toric variety coincides with the sheaf of reflexive differentials ([GNAPGP88, V.4]).
- domain assumption Weber's theorem: ker(H^{2p}(X,Q) -> IH^{2p}(X,Q)) = W_{2p-1} H^{2p}(X,Q) ([PP24, Rem. 6.5]).
- domain assumption Toric varieties are normal with rational singularities, and structure sheaves of proper toric varieties are RGamma-acyclic ([CLS11, Thm. 9.2.5]).
- domain assumption Hard Lefschetz theorem for intersection cohomology of projective toric varieties.
- domain assumption Results from the authors' previous preprint [KV25]: H^{2p}(X,Q) is mixed of Hodge-Tate type; the last map of Ish^p_X is surjective for p >= n/2; lcdef equals m for (m+3)-dimensional non-simplicial isolated toric varieties.
Cite this review
Pith. "Pith review of Lefschetz morphisms on singular cohomology and local cohomological dimension of toric varieties." pith.science (2026). https://pith.science/paper/MITZ6GM7
@misc{pith2026250603026,
author = {Pith},
title = {Pith review of: Lefschetz morphisms on singular cohomology and local cohomological dimension of toric varieties},
year = {2026},
howpublished = {\url{https://pith.science/paper/MITZ6GM7}},
note = {Machine review of arXiv:2506.03026}
}
read the original abstract
Given a proper toric variety and a line bundle on it, we describe the morphism on singular cohomology given by the cup product with the Chern class of that line bundle in terms of the data of the associated fan. Using that, we relate the local cohomological dimension of an affine toric variety with the Lefschetz morphism on the singular cohomology of a projective toric variety of one dimension lower. As a corollary, we show that the local cohomological defect is not a combinatorial invariant. We also produce numerous examples of toric varieties in every dimension with any possible local cohomological defect, by showing that the local cohomological defect remains unchanged under taking a pyramid.
Reference graph
Works this paper leans on
-
[1]
[BM71] H. Bruggesser and P. Mani,Shellable decompositions of cells and spheres, Math. Scand.29 (1971), 197–205. MR328944↑8 [CLS11] David A. Cox, John B. Little, and Henry K. Schenck,Toric varieties, Graduate Studies in Mathematics, vol. 124, American Mathematical Society, Providence, RI,
work page 1971
-
[22]
MR4202748↑11, 12 [Web04] Andrzej Weber,Pure homology of algebraic varieties, Topology43(2004), no. 3, 635–644. MR2041634↑17 24 HYUNSUK KIM AND SRIDHAR VENKATESH Department of Mathematics, University of Michigan, 530 Church Street, Ann Arbor, MI 48109, USA Email address:kimhysuk@umich.edu Department of Mathematics, University of Michigan, 530 Church Street...
work page 2004
-
[58]
MR4491455↑2, 3 [PP24] Sung Gi Park and Mihnea Popa,Lefschetz theorems, Q-factoriality, and Hodge symmetry for singular varieties, arXiv preprint arXiv:2410.15638 (2024).↑17 [PS24] Mihnea Popa and Wanchun Shen,Du Bois complexes of cones over singular varieties, local cohomological dimension, and K-groups, arXiv:2406.03593 (2024).↑1 [PS73] C. Peskine and L....
arXiv 2024
-
[2005]
An intro- duction. MR2093043↑12 [Ish87] Masa-Nori Ishida,Torus embeddings and de Rham complexes, Commutative algebra and com- binatorics (Kyoto, 1985), 1987, pp. 111–145. MR951199↑1, 2, 8, 9 [KV25] Hyunsuk Kim and Sridhar Venkatesh,Local cohomology and singular cohomology of toric vari- eties via mixed Hodge modules, arXiv preprint arXiv:2505.10417 (2025)...
arXiv 2025
-
[2011]
MR2810322↑4, 6, 7, 10, 11 [DB81] Philippe Du Bois,Complexe de de Rham filtr´ e d’une vari´ et´ e singuli` ere, Bull. Soc. Math. France 109(1981), no. 1, 41–81. MR613848↑5 [DOR25] Bradley Dirks, Sebasti´ an Olano, and Debaditya Raychaudhury,A Hodge theoretic generalization ofQ-homology manifolds, arXiv preprint arXiv:2501.14065 (2025).↑22 [DT16] Hailong Da...
arXiv 1981
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.