A Hodge Theoretic Generalization of $\mathbb{Q}$-Homology Manifolds II: Local Complete Intersections

Bradley Dirks; Debaditya Raychaudhury; Sebastian Olano

REVIEW 3 major objections 4 minor 3 references

A Hodge Theoretic Generalization of $\mathbb{Q}$-Homology Manifolds II: Local Complete Intersections

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read For hypersurfaces, HRH equals the first Hodge-filtration jump of unipotent vanishing cycles, shifted by n−2; for complete intersections that same jump bounds HRH, with the gap measured by 'missed jumps'.

desk verdict Genuinely new comparison inequalities for HRH on local complete intersections, but the results are built on the unpublished [DOR26]; the math looks coherent and deserving of refereeing, not desk rejection. read the letter →

arxiv 2607.25861 v1 pith:XOL52SXS submitted 2026-07-28 math.AG

Bradley Dirks , Sebastian Olano , Debaditya Raychaudhury This is my paper · ORCID

classification math.AG MSC 14B0514F1032S35

keywords HodgerationalhomologymanifoldlocalcompleteintersectionBernstein–SatopolynomialspectrumV-filtrationVerdierspecializationvanishingcyclesminimalexponent

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

HRH(Z) is a proposed Hodge-theoretic refinement of the rational homology manifold condition: instead of asking whether the link has the homology of a sphere, it asks how deep into the Hodge filtration on local cohomology a nontrivial class survives. The paper proves that in the hypersurface case this level is completely pinned down by classical invariants: HRH_x(Z) equals the first Hodge filtration jump of the unipotent vanishing cycles φ_{f,1}(O_X) shifted by n−2, and in isolated singularities this equals the smallest integer spectral number minus two. For local complete intersections the authors show that the same jump, computed instead on the unipotent Verdier specialization module Q_Z, gives a lower bound HRH(Z) ≥ p(Q_Z,F)+n−1 and that the DMS spectrum gives HRH_x(Z) ≥ Sp_min,Z(Z,x)−2 in the isolated case. The gap between these bounds is controlled by a count of 'missed jumps' in the Hodge filtrations of the monodromic pieces of Q_Z, and the vanishing Q_Z = 0 is characterized by the rational homology manifold property of the generic linear combination hypersurface.

What carries the argument

The engine of the paper is the V-filtration of Kashiwara–Malgrange on the graph-embedded structure module B_f = Γ_*(Q_X^H[n]), and its associated Verdier specialization Sp(B_f) along Z ⊆ X. The module Q_Z is the cokernel in 0 → L → Sp(B_f)^Z → Q_Z → 0, where L is the trivial Hodge module on Z × A^r; its first Hodge filtration jump p(Q_Z,F) is the key numerical invariant. A secondary but equally important object is the number j(Z) of 'missed jumps' — the count of ℓ ∈ [1,r−1] for which p(Q_{r−ℓ}, F) = p(Q_{r−ℓ+1}, F) — which measures when the Hodge filtrations on consecutive monodromic pieces align, and which turns the general inequalities into equalities.

What would settle it

For a cone over a smooth complete intersection of degree (a_1,...,a_r) in P^{n-1}, the Milnor-fiber spectral numbers and link Hodge numbers are known; check whether HRH computed from the V-filtration (via the hypersurface formula) satisfies HRH = Sp_min − 2. A mismatch in any such example would refute the isolated-case equalities. More directly, if an isolated LCI singularity is found where HRH_x(Z) < Sp_min,Z(Z,x) − 2, then Theorem C fails.

Watch

Extended reading notes

Core claim

The central discovery is that the invariant HRH, defined as the largest k for which F^{k−n}W_{n+1}H^1_Z(O_X) = F^{k−n}H^1_Z(O_X), is completely determined by the Hodge filtration of unipotent nearby/vanishing cycles in the hypersurface case: HRH_x(Z) = p(φ_{f,1}(O_X)_x,F) + n − 2 (Theorem A); if x is isolated this equals Sp_min,Z(Z,x) − 2. In codimension r > 1, the same role is played by the unipotent Verdier specialization module Q_Z, giving p(Q_Z,F)+n−1 ≤ HRH(Z) and, for isolated singularities, HRH_x(Z) ≥ Sp_min,Z(Z,x) − 2 (Theorems B and C). Equality is controlled by the absence of 'missed jumps' among the Hodge filtrations on the monodromic pieces Q_k. The vanishing Q_Z = 0 is characteri

Load-bearing premise

The paper relies on an unpublished companion paper for the definition of HRH and, crucially, for the equivalence (Corollary G there) between HRH ≥ k and the vanishing of the link Hodge numbers ℓ_{p,q} used in Proposition 6.1; if that equivalence is wrong or unavailable, Theorem C and the spectrum-to-HRH inequalities lose their support.

Editorial extensions

If this is right

For any hypersurface singularity, HRH_x(Z) equals the first Hodge jump of unipotent vanishing cycles shifted by n−2; in particular, the reduced Bernstein–Sato polynomial determines HRH_x(Z) via eαbar_{Z,x}(f).
In isolated hypersurface singularities, HRH_x(Z) = Sp_min,Z(Z,x) − 2, so the classical Steenbrink spectrum fully determines the Hodge rational homology level.
For local complete intersections, HRH(Z) ≥ p(Q_Z,F) + n − 1; thus the first jump on the unipotent Verdier specialization forces a quantitative bound on how close Z is to a rational homology manifold.
The equality conditions in the bounds are governed by the missed-jump count j(Z); when j(Z) = 0 the invariants HRH(Z), p(Q_Z,F), and HRH(g|_U) differ only by explicit shifts in r.
If Z has rational singularities and eαZ(Z) = +∞ (i.e., the reduced Bernstein–Sato polynomial has no integer roots), then Z is a rational homology manifold; the converse fails, as Torrelli's example shows.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The missed-jump count j(Z) likely measures the failure of the Hodge filtration on Q_Z to be 'as spread out as possible'; if so, it should be computable from the dimensions s_p of the Milnor-fiber Hodge pieces, and one could test whether j(Z) equals the number of p < r for which s_{d−p} = s_p in isolated singularities.
The general linear combination trick opens a practical route to compute HRH for complete intersections: pick a generic linear combination g, compute the spectrum (or Bernstein–Sato polynomial) of g|_U, and correct by r and j(Z). This could turn HRH into a computable invariant for concrete families, e.g., cones over smooth complete intersections.
Conjecture 5.20, if true, would give a purely combinatorial criterion: Q_Z = 0 (and hence HRH maximal) can be read off from whether the Bernstein–Sato polynomial of the tuple has extra factors beyond (s+r); testing this on the Thorrelli-type product examples would be a concrete check.
The distinction the paper draws between Q_Z = 0 and rational-homology-manifold-ness suggests that in codimension > 1 the Hodge-theoretic boundary is not the ordinary link condition but a finer one; one might expect HRH to relate to the vanishing of the Du Bois complex or to k-rationality thresholds in a way that Q_Z does not.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Jason — quick take on arXiv:2607.25861. The paper gives a real new dictionary: for hypersurfaces, HRH_x(Z) is sandwiched between the reduced Bernstein–Sato integer root and the minimal integer spectral number, with equality of the second bound in the isolated case. The LCI inequalities involving p(Q_Z,F), the DMS spectrum, and the unipotent Verdier specialization module Q_Z are new, and the characterization of Q_Z=0 via the general linear combination g|_U being a rational homology manifold is a clean statement. The 'missed jumps' language in Section 5 is a useful way to organize the comparisons. The internal arguments read as coherent; I did not find an obvious contradiction.

The soft spots are real but not disqualifying. First, the central invariant HRH is not defined in this paper; Section 3 starts by recalling [DOR26, Thm. D(1)] to identify it, and Proposition 6.1/Theorem C use [DOR26, Cor. G] twice to pass between HRH and vanishing of link Hodge numbers. If that corollary is missing or wrong, the spectrum-to-HRH inequalities lose their support. That is a load-bearing external dependency, and it is not a matter of style. Second, the proof of Theorem A contains an 'easy exercise' and an 'induction on k' on exactly the filtered weight-filtration formula that the argument turns on. Both steps are probably correct—I suspect an expert can fill them in—but they are compressed enough that a referee will need to ask for details. Third, the one Macaulay2 computation (Example 5.21) is stated without attached code or even a certificate; minor, but easy to fix.

On the other hand, Section 5 does define and prove the non-HRH invariants from scratch, and the reliance on same-group papers like [Dir23] and [CDMO24] is not circular in a bad way—those hold established results. The stress-test note about [DOR26] is on target. This is a paper for singularity theorists and Hodge theory people. If Part I appears and the compressed proofs are expanded, the results will be worth citing. As it stands, I would not build on it yet, but it deserves a serious referee. Recommendation: send to peer review, with explicit instruction to the authors to either include a full proof of Cor. G or restructure so the paper is self-contained enough to be evaluated.

Referee Report

3 major / 4 minor

Summary. The paper studies the invariant HRH ("Hodge rational homology" manifold level) for local complete intersection subvarieties of smooth complex varieties. In the hypersurface case, Theorem A asserts for each x in Z that eαbar_{Z,x}(f)-2 ≤ HRH_x(Z) = p(φ_{f,1}(O_X)_x,F)+n-2 ≤ Sp_{min,Z}(Z,x)-2, with equality of the second inequality at isolated singularities. For general local complete intersections, Theorem B bounds HRH(Z) below by p(Q_Z,F)+n-1 and bounds the integral spectrum below by p(Q^Z_x,F)+n+1; Theorem C gives HRH_x(Z) ≥ Sp_{min,Z}(Z,x)-2 for isolated LCI singularities; Theorem D characterizes vanishing of Q_Z in terms of the rational homology manifold property of the general linear combination hypersurface V(g|_U). The proofs use the V-filtration, Verdier specialization, Fourier-Laplace transforms, and results from the authors' companion paper [DOR26], which is listed as "In preparation".

Significance. If the main results are correct, they give a striking Hodge-theoretic description of rational homology manifold behavior: for hypersurfaces, HRH is completely determined by the first Hodge-filtration jump of unipotent vanishing cycles, hence by the DMS spectrum; for complete intersections, the paper provides new inequalities linking HRH, the minimal integer spectral number, and the reduced Bernstein-Sato polynomial. The paper contains no fitted parameters, and the comparisons are genuine inequalities rather than definitions in disguise. The explicit examples, including the Thom-Sebastiani examples and a Macaulay2 computation in Example 5.21, are useful. However, the central claims are currently conditional on the unpublished companion paper [DOR26], which supplies the definition of HRH and the key link-Hodge-number criterion used in Proposition 6.1 and Theorem C; this is a substantial verification gap that must be addressed before the results can be considered established.

major comments (3)

[§3 and §6 (HRH via [DOR26])] The invariant HRH is never defined in this paper. Section 3 identifies k-Hodge rational homology with the filtered equality F^{k-n}W_{n+1}H^1_Z(O_X)=F^{k-n}H^1_Z(O_X) by quoting [DOR26, Thm. D(1)], and Proposition 6.1 uses [DOR26, Cor. G] to translate HRH(Z)≥k into vanishing of link Hodge numbers ℓ_{p,q}. Since [DOR26] is listed as "In preparation", the main comparisons to HRH are conditional on an external document. Please state the precise needed results from [DOR26] or include proofs, so that the paper is self-contained on this load-bearing point.
[§3, proof of Theorem A] The proof of Theorem A is too compressed at its most crucial step. The filtered weight-filtration formula F^pW_{n-1+i}Gr^1_V(B_f)=Σ_{ℓ≥max{0,-i}}(t∂_t)^ℓ F^{p-ℓ}ker((t∂_t)^{1+i+2ℓ}) is said to be "an easy exercise" following strictness, and the second half of the proof invokes "By induction on k" without spelling out the induction parameter or the base case. Since Theorem A is the central hypersurface characterization, these steps need a complete and detailed proof.
[§6, Proposition 6.1 and Theorem C] Proposition 6.1 is the bridge from the Milnor-fiber numbers s_p to HRH, and Theorem C depends entirely on it. Its proof invokes [DOR26, Cor. G] in both directions, but the exact statement of that corollary is not given. Even accepting [DOR26], the induction in the converse direction is only sketched: after proving s_{d-p}-s_p=0 for p≤k, the implication ℓ_{d-k,k}=0 ⇒ HRH(Z)≥k requires precisely the missing form of Cor. G. This is a load-bearing gap, not a presentation issue.

minor comments (4)

[Plan of the paper] The plan refers to "Section B", "Section C", and "Section D", while the actual sections are numbered 1-8; use consistent numbering.
[Remark 5.15] The inequality HRH(Z)<+∞ ⇒ HRH(Z)≤(d−3)/2 is stated without reference or proof. Add a citation or a short justification.
[Example 5.21] The Macaulay2 computation would be more reproducible if the code and version were included or mentioned in a footnote.
[§6, Proposition 6.4] The notation D(Q_{≠Z}) in the proof is not defined; clarify that it means the graded-dual construction on the monodromic pieces.

Circularity Check

2 steps flagged · score 4.0 of 10

No constructional circularity: The inequalities compare independently defined invariants and no fitted input is relabeled as a prediction. The main circularity burden is a load-bearing self-citation: HRH and its key characterization are imported from the same authors' unpublished [DOR26], especially in Proposition 6.1 and Theorem C.

self citation load bearing [Section 3, proof of Theorem A; also References [DOR26]]
"Recall that by [DOR26, Thm. D(1)], Z is k-Hodge rational homology if F^{k-n}W_{n+1}H^1_Z(O_X) = F^{k-n}H^1_Z(O_X)."

The HRH-side of the paper's main hypersurface theorem is identified with a V-filtration condition by citing [DOR26, Thm. D(1)] rather than by a proof given in this paper. [DOR26] is by the same three authors and is listed in the references as 'In preparation.' This makes the central comparison conditional on a same-author unpublished statement. It is not a constructional reduction (no parameter is fitted, and HRH is not defined here through Sp_min or p(Q,F)), but it is load-bearing self-citation.
self citation load bearing [Section 6, proof of Proposition 6.1]
"Suppose HRH(Z)≥k. Then by [DOR26, Cor. G], 0 = ℓ_{d-p,p} = ℓ_{p,d-p-1} for p≤k, and by combining (6.2), (6.3), and using induction on k we get s_{d-p} −s_p = 0 for all 0≤p≤k."

Proposition 6.1 is the critical bridge used to prove Theorem C: it converts vanishing of the spectral differences s_{d-p}−s_p into the statement HRH(Z)≥k. Both directions of this conversion invoke [DOR26, Cor. G], a result from the authors' unpublished predecessor. If Cor. G is unavailable or false, Theorem C loses its support. This is a verification dependency on a self-citation rather than an internal derivation from definitions given in this paper.

full rationale

The paper does not derive HRH from the comparison invariants: HRH is introduced in [DOR26], while the V-filtration, Verdier specialization, spectrum, and p(Q,F) are set up independently here or in published work. No parameters are fitted to data and then called predictions; Theorem A, Theorem B, and Theorem D are genuine inequalities between distinct invariants. The main concern is the repeated load-bearing use of [DOR26, Thm. D(1)] and [DOR26, Cor. G] to identify HRH with filtered conditions on local cohomology and link Hodge numbers. Because [DOR26] is by the same authors and marked 'In preparation,' it is not machine-checked or externally available evidence. However, this is not a case where an equation reduces by construction to its input: the V-filtration and spectrum computations in Section 5 are carried out in this paper, and the inequalities would be meaningful once the Part I foundation is supplied. Accordingly, the circularity score is moderate: the central claim still has independent content, but a key step is supported by a same-author unpublished citation.

Assumptions & free parameters 0 free parameters · 5 assumptions · 1 invented entities

The paper rests on Saito's mixed Hodge module machinery and on several cited results, many authored by the same group; the distinguishing load is the unpublished predecessor DOR26 defining HRH and proving Cor. G. No free parameters are fitted.

assumptions (5)

domain assumption Existence and strictness properties of the V-filtration and Hodge filtration for mixed Hodge modules (Saito's theory; Propositions 1.2, 1.5, 1.6 from CD23/CDS23).
Used throughout §2–§5 to construct Sp(B_f), filtered acyclicity, and p(Q,F) comparisons; accepted as published theorems but not re-proved here.
ad hoc to paper HRH is a well-defined invariant satisfying [DOR26, Thm. D(1)] and [DOR26, Cor. G] in the link-Hodge-number criterion.
The paper cites [DOR26] as 'In preparation'; Cor. G is load-bearing in Prop 6.1 and Theorem C and is not stated or proved in this text.
domain assumption Formulas (6.2) s_{d-p}-s_p = ℓ_{p,d-p-1}-ℓ_{p,d-p} and (6.3) of Friedman–Laza identify spectral-number jumps with link Hodge numbers in isolated LCI singularities.
Imported from [FL24, Prop. 2.11/2.12] and used for Theorem C and Prop 6.1.
domain assumption Bernstein–Sato polynomial factorizations for tuples: b_f(s)= e b_g(s)= e b_{g|U}(s)∏_{i∈I}(s+r+i), and Thom-Sebastiani product formulas.
Imported from [Mus22], [BMS06], [Lee24], [Dir23]; used in Cor 5.14, Conjecture 5.20, and Examples 7.2.
domain assumption Macaulay2 computation in Example 5.21 gives b_g(s)=b_{g|U}(s) with the displayed factorization.
Computational evidence for a case where the factor set I is empty; script and version are not provided.

invented entities (1)

HRH(Z), the Hodge rational homology manifold level
purpose: Measures the first Hodge level at which a singular variety fails to be a rational homology manifold; it is the central target invariant of the paper.
Introduced in the authors' unpublished [DOR26]. This paper compares it with known invariants, but no independent external computation or formal verification of HRH is provided outside the authors' program.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Hodge Theoretic Generalization of $\mathbb{Q}$-Homology Manifolds II: Local Complete Intersections." pith.science (2026). https://pith.science/paper/XOL52SXS

@misc{pith2026260725861,
  author       = {Pith},
  title        = {Pith review of: A Hodge Theoretic Generalization of $\mathbbQ$-Homology Manifolds II: Local Complete Intersections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XOL52SXS}},
  note         = {Machine review of arXiv:2607.25861}
}

abstract

Recently, the authors introduced and studied a singularity invariant of a complex algebraic variety $Z$, written $\HRH(Z)$ (for ``Hodge rational homology'' manifold level). In this paper, we focus on local complete intersection subvarieties. We relate $\HRH(Z)$ to various well-known invariants, like Bernstein--Sato polynomials and the Dimca-Maisonobe-Saito spectrum. In the hypersurface case it turns out that $\HRH(Z)$ can be completely characterized by these invariants, though higher codimension case is more subtle.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references · 2 linked inside Pith

[1]

Math.142(2006), no

[BMS06] Nero Budur, Mircea Mustat ¸ˇ a, and Morihiko Saito,Bernstein-Sato polynomials of arbitrary varieties, Com- pos. Math.142(2006), no. 3, 779–797. MR2231202↑3, 6, 18, 24 [CD23] Qianyu Chen and Bradley Dirks,OnV-filtration, Hodge filtration and Fourier transform, Selecta Math. (N.S.)29(2023), no. 4, Paper No. 50,

2006
[28]

MR4603110↑9 [PP25a] Sung Gi Park and Mihnea Popa,Hodge symmetry and Lefschetz theorems for singular varieties, preprint arXiv: (2025).↑1 [PP25b] ,Q-factoriality and Hodge-Du Bois theory, preprint arXiv:2508.17748 (2025).↑1 [PSV24] Mihnea Popa, Wanchun Shen, and Anh Duc Vo,Injectivity and Vanishing for the Du Bois Complexes of Isolated Singularities, arXiv...

arXiv 2025
[76]

MR4610959↑5, 6 [CDM22] Qianyu Chen, Bradley Dirks, and Mircea Mustat ¸˘ a,The minimal exponent andk-rationality for local com- plete intersections, arXiv e-prints (December 2022), arXiv:2212.01898, available at2212.01898. to appear in Journal de l’ ´Ecole Polytechnique.↑1, 3 [CDMO24] Qianyu Chen, Bradley Dirks, Mircea Mustat ¸˘ a, and Sebasti´ an Olano,V-...

arXiv 2022

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

Math.142(2006), no

[BMS06] Nero Budur, Mircea Mustat ¸ˇ a, and Morihiko Saito,Bernstein-Sato polynomials of arbitrary varieties, Com- pos. Math.142(2006), no. 3, 779–797. MR2231202↑3, 6, 18, 24 [CD23] Qianyu Chen and Bradley Dirks,OnV-filtration, Hodge filtration and Fourier transform, Selecta Math. (N.S.)29(2023), no. 4, Paper No. 50,

2006

[2] [28]

MR4603110↑9 [PP25a] Sung Gi Park and Mihnea Popa,Hodge symmetry and Lefschetz theorems for singular varieties, preprint arXiv: (2025).↑1 [PP25b] ,Q-factoriality and Hodge-Du Bois theory, preprint arXiv:2508.17748 (2025).↑1 [PSV24] Mihnea Popa, Wanchun Shen, and Anh Duc Vo,Injectivity and Vanishing for the Du Bois Complexes of Isolated Singularities, arXiv...

arXiv 2025

[3] [76]

MR4610959↑5, 6 [CDM22] Qianyu Chen, Bradley Dirks, and Mircea Mustat ¸˘ a,The minimal exponent andk-rationality for local com- plete intersections, arXiv e-prints (December 2022), arXiv:2212.01898, available at2212.01898. to appear in Journal de l’ ´Ecole Polytechnique.↑1, 3 [CDMO24] Qianyu Chen, Bradley Dirks, Mircea Mustat ¸˘ a, and Sebasti´ an Olano,V-...

arXiv 2022