REVIEW 1 major objections 4 minor 20 references
Hodge-Grothendieck classes and monodromy invariants of nearby cycles sheaves
T0 review · 1 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read The Hodge-Grothendieck class of unipotent nearby cycles depends only on the reduced central fibre.
desk verdict A short, honest note: the Hodge-Grothendieck class of unipotent nearby cycles really is intrinsic to the reduced central fiber, and the proof is a clean corollary of Saito; the one real gap is an unproved reduction in the proof of Theorem 4.2 that is fillable with standard facts. 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 load-bearing mechanism is the weight/monodromy filtration on $p\Psi_f^u(M)$, governed by the logarithm of unipotent monodromy $N:\psi_f^u(M)\to\psi_f^u(M)(-1)$. For pure $M$ of weight $n$, the associated graded pieces of $p\Psi_f^u(M)$ are reconstructed from graded pieces of $\ker(N)$ by powers of $N$, and $\ker(N)$ is canonically isomorphic to $i_*j_{!*}M[-1]$. Because the filtration induced on the intermediate extension $i_*j_{!*}M$ is independent of $f$, the Grothendieck class computed from these graded pieces is also independent of $f$.
What would settle it
A direct local check: take $X=\mathbb{A}^1$, $f=t^2$ and $g=t^3$ (both have reduced fibre $\{0\}$), and take $M$ to be the skyscraper sheaf at $0$. The theorem and its proof predict $[\psi_f^u(M)]=[\psi_g^u(M)]=0$ because $M$ is supported on the central fibre; if an explicit computation of the two nearby-cycle complexes gives a nonzero class, or differs between $f$ and $g$, the key reduction in the proof of Theorem 4.2 fails.
Extended reading notes
Core claim
The central claim is Theorem 4.2: let $f,g:X\to \mathbb{A}^1$ have $f^{-1}(0)_{\mathrm{red}}=g^{-1}(0)_{\mathrm{red}}$. Then in the Grothendieck group $K_0(X_0)$, $[\psi_f^u(M)]=[\psi_g^u(M)]$ for every $M$ in $D(X)$. The proof restricts to pure objects $M$ supported on the complement $X^*$ of the central fibre, and expresses the class of $p\Psi_f^u(M)=\psi_f^u(j_*M)[-1]$ as a sum over graded pieces of the weight filtration of the intermediate extension $i_*j_{!*}M$, with Tate twists. Since the filtration induced on $i_*j_{!*}M$ is independent of the defining function, the class is equation-independent. The paper also proves Theorem 5.2: for proper $f$ and pure $M$, the sequence $H^k(X_0;i^*M) \to H^k(X_0;\psi_f^u(M)) \xrightarrow{N} H^k(X_0;\psi_f^u(M))(-1)$ is exact for each $k$,
Load-bearing premise
The proof begins by declaring that it suffices to verify the equality for pure objects supported on the complement $X\setminus X_0$; this reduction is stated without proof and carries the whole argument.
Editorial extensions
If this is right
- If f and g define the same reduced central fibre, then [ψ^u_f(M)]=[ψ^u_g(M)] in K_0(X_0) for every object M of the derived category.
- The proof gives an explicit formula for the unipotent nearby-cycles class in terms of the intermediate extension of M and Tate twists, so the equation enters only through the choice of the reduced fibre.
- For proper f and pure M, the generalized local invariant cycles sequence through N is exact in every degree, extending the classical statement beyond constant sheaves.
- In the rationally smooth case with M the constant Hodge module, the generalized statement recovers the classical local invariant cycles theorem.
- For ordinary constructible sheaves, the same invariance follows without Hodge theory by working with the monodromy filtration and the independence of Verdier specialization.
Reading between the lines
- Extrapolating from the class-level result, one could define the unipotent nearby-cycles Hodge-Grothendieck class for any reduced closed subscheme without choosing a defining equation, as the common value over all equations with that reduced fibre; the paper does not propose such a definition.
- The theorem proves equality of Grothendieck classes, not isomorphism of functors; a natural stricter question left open here is whether ψ^u_f(M) and ψ^u_g(M) are actually isomorphic when the reduced fibres agree.
- Because the proof uses only the weight formalism and the canonical N-triangle, the same invariance should hold in any setting with an analogous weight package, such as étale ℓ-adic sheaves with the standard axioms; the paper states only the Hodge-module and topological-sheaf versions.
- The argument for Theorem 5.2 is phrased for proper maps and pure objects; a reader might ask whether exactness persists for non-proper f when the pure object is replaced by one with appropriate support conditions.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This short note studies unipotent nearby cycles in the category of mixed Hodge modules. Its main result, Theorem 4.2, asserts that for two morphisms f,g: X → A^1 with the same reduced zero fibre, the induced maps [ψ^u_f] and [ψ^u_g] on K_0(X_0) coincide, i.e. the Hodge–Grothendieck class of unipotent nearby cycles is intrinsic to the reduced central fibre. The proof combines Saito's primitive-decomposition formula (3.3.1) with Proposition 2.3, which identifies ker(N) with i^* j_{!*} M[-1]. A second result, Theorem 5.2, is a generalized local invariant cycles exact sequence for proper f, deduced from weight arguments. The paper is candid that it makes no claims to originality and relies on external theorems of Saito and Deligne.
Significance. If the gaps noted below are filled, this would be a clean and useful structural fact: the Hodge–Grothendieck class of the unipotent part of nearby cycles is independent of the defining equation, directly addressing Williamson's question in a Hodge-theoretic context. The reliance on established machinery is explicit and the argument is not circular: the target independence claim is not used as an input. The note is concise, honest, and the main idea—that the primitive part is controlled by i^* j_{!*} M—is sound. The main issue is that the proof of Theorem 4.2 as written omits a load-bearing reduction from arbitrary objects of D(X) to pure objects of M(X^*).
major comments (1)
- [Section 4, proof of Theorem 4.2] The proof opens with 'It suffices to show [pΨ^u_f(M)] = [pΨ^u_g(M)] for M∈M(X^*)' and then 'We may also assume M is pure of weight n'. Neither reduction is justified. For arbitrary M∈D(X), one must use the triangle i_* i^! M → M → j_* j^* M → and prove that ψ^u_f(i_* i^! M)=0; this is true because the Milnor fibre lies in X\X_0, so any object supported on X_0 has zero nearby cycles. One must then pass from ψ^u_f(j_* j^* M) to pΨ^u_f(j^* M) and use K_0(D(X^*)) ≅ K_0(M(X^*)). The purity reduction needs additivity over the weight filtration of M. These are standard facts, but they are load-bearing for the theorem as stated for all M∈D(X), and they are not mentioned. Please add a short lemma or paragraph supplying these steps.
minor comments (4)
- [Proposition 2.3(i)] The statement writes 'ker(N) ≃ i_* j_{!*} M[-1]', but the right-hand side should be an object of M(X_0), so the intended functor is i^*, not i_*. The proof itself uses i^* j_{!*} M. Please correct the display to i^* j_{!*} M[-1].
- [§5.1 and proof of Theorem 5.2] The same i_*/i^* ambiguity appears in the maps 'i_* M → i_* j_* j^* M → ψ^u_f(M)' and in the displayed exact sequence 'H^k(X_0; i_* M) → ...'. Since these are cohomology groups on X_0, the functors should be i^*. Please clarify the notation throughout; using a single symbol i* for i_*, i^*, and i^! is confusing.
- [Remark 4.3] The phrase 'induced filtration on i_* j_{!*} M' should presumably be 'on i^* j_{!*} M'. Also, there is a typo 'orginality' in the introduction.
- [Section 4, Remark 4.3] The remark claims that in the topological setting 'the only extra ingredient needed' is the independence of the filtration on i^* j_{!*} M. This is accurate, but the phrase 'the induced filtration' could be expanded: one needs to specify that it is the monodromy filtration induced from nearby cycles, and that [ELM] supplies the independence from f. A one-sentence elaboration would help.
Circularity Check
No significant circularity: the independence claim is derived from Saito's primitive decomposition and the intermediate extension, neither of which assumes the target result.
full rationale
Theorem 4.2's claim — that the Grothendieck class of unipotent nearby cycles depends only on the reduced central fiber — is not circular. The proof computes [pΨ^u_f(M)] for pure M∈M(X*) via the primitive decomposition (3.3.1) and Proposition 2.3(i), which identifies ker(N) with i_* j_{!*}(M)[-1]. The intermediate extension j_{!*}M and its weight filtration depend only on the open complement X* = X∖X0, i.e. only on the reduced central fiber, not on the particular equation f. The resulting class is therefore manifestly independent of f. Saito's results, Deligne's invariant cycles theorem, and [ELM] are external, published mathematical inputs; they do not contain Theorem 4.2 as an assumption. There are no fitted parameters, no data-dependent quantities, and no self-citations. The only rigor issue in the proof of Theorem 4.2 is the unmotivated opening reduction: 'It suffices to show [pΨ^u_f(M)] = [pΨ^u_g(M)] for M∈M(X*).' This step is not proved in the text; it needs the standard vanishing of ψ^u_f on objects supported on X0 and the fact that K0(D(X*)) is generated by classes from M(X*). That is an omitted proof of a standard, fillable lemma, not a circular step: it does not assume the conclusion and does not identify the target with an input by definition. Similarly, comparing K0(X0) for possibly different nilpotent structures on the same reduced fiber is a technical point left implicit, but again not circular. Overall the derivation is self-contained modulo standard Hodge-module facts, and the independence result has genuine content.
Assumptions & free parameters
assumptions (6)
- domain assumption Existence and six-functor formalism of mixed Hodge modules on varieties over C, with a finite increasing weight filtration W and strict compatibility of morphisms with W.
- standard math The unipotent nearby cycles functor ψ^u_f satisfies the canonical distinguished triangle (2.2.1) and the weight property N W_k ⊂ W_{k-2}(-1).
- standard math Primitive decomposition identity (3.3.1): for M pure of weight n, Gr^W_k pΨ^u_f(M) is a direct sum of shifted graded pieces of ker(N).
- standard math Intermediate extension properties: for M∈M(X^*), the intermediate extension j_{!*}M satisfies pH^k(i^*j_{!*}M)=0 for k≠-1 and pH^0(i^!j_{!*}M)=0.
- standard math Proper pushforward preserves weights.
- standard math For f proper, the weight filtration on H^k(X_0;ψ^u_f(M)) is the monodromy filtration centered at n+k.
Cite this review
Pith. "Pith review of Hodge-Grothendieck classes and monodromy invariants of nearby cycles sheaves." pith.science (2026). https://pith.science/paper/XA7RWAML
@misc{pith2026250900984,
author = {Pith},
title = {Pith review of: Hodge-Grothendieck classes and monodromy invariants of nearby cycles sheaves},
year = {2026},
howpublished = {\url{https://pith.science/paper/XA7RWAML}},
note = {Machine review of arXiv:2509.00984}
}
read the original abstract
Remarks on the Hodge-Grothendieck class of the nearby cycles functor and a generalized local invariant cycles result.
Reference graph
Works this paper leans on
-
[1]
Brylinski , Transformations canoniques, dualit\'e projective, th\'eorie de Lefschetz, transformations de Fourier et sommes trigonom\'etriques , Ast\'erisque 140-141(1986)
J-L. Brylinski , Transformations canoniques, dualit\'e projective, th\'eorie de Lefschetz, transformations de Fourier et sommes trigonom\'etriques , Ast\'erisque 140-141(1986)
1986
-
[2]
Brylinski, B
J-L. Brylinski, B. Malgrange, J-L. Verdier , Transformation de Fourier g\'eom\'etrique II , C. R. Acad. Sci. Paris 303 (1986)
1986
-
[3]
Bott, L.W
R. Bott, L.W. Tu , Differential Forms in Algebraic Topology , GTM 82, Springer-Verlag (1982)
1982
-
[4]
J. Bernstein, I.M. Gelfand, S.I. Gelfand , A certain category of -modules , Functional Analysis and its Applications 10, no. 2 (1976), 1-8
work page 1976
-
[5]
Deligne , Th\'eorie de Hodge III , Publ
P. Deligne , Th\'eorie de Hodge III , Publ. I.H.\'E.S 44 (1974), 5-77
1974
-
[6]
Deligne , La conjecture de Weil: II , Publ
P. Deligne , La conjecture de Weil: II , Publ. I.H.\'E.S 52 (1980)
1980
-
[7]
Bernstein, V
J. Bernstein, V. Lunts , Equivariant sheaves and functors , Springer-Verlag, Berlin (1994)
1994
-
[8]
Bernstein, V
J. Bernstein, V. Lunts , Equivariant sheaves and functors , Lecture Notes in Math. 1578, Springer-Verlag, Berlin (1994)
1994
Show all 20 references
-
[9]
El Zein, D.T
F. El Zein, D.T. L\^e, L. Migliorini , A topological construction of the weight filtration , Manuscripta Math. 133 (2010)
2010
-
[10]
Kashiwara, P
M. Kashiwara, P. Schapira , Categories and Sheaves , Springer-Verlag, Berlin (1990)
1990
-
[11]
Langlands , On the Classification of Irreducible Representations of Real Algebraic Groups , preprint I.A.S
R.P. Langlands , On the Classification of Irreducible Representations of Real Algebraic Groups , preprint I.A.S. (1973), appeared in Math. Surveys and Monographs 31, AMS (1988)
1973
-
[12]
Serre , Espaces fibr\'es alg\'ebriques , Anneaux de Chow et Applications, Se\'eminaire C
J.-P. Serre , Espaces fibr\'es alg\'ebriques , Anneaux de Chow et Applications, Se\'eminaire C. Chevalley, E.N.S., Paris (1958)
1958
-
[13]
SGA 7, volume II, Expos\'e XIII
-
[14]
Saito , Modules de Hodge polarisables , Publ
M. Saito , Modules de Hodge polarisables , Publ. RIMS 24 (1988)
1988
-
[15]
Saito , Introduction to mixed Hodge modules , Ast\'erisque 179-180 (1989)
M. Saito , Introduction to mixed Hodge modules , Ast\'erisque 179-180 (1989)
1989
-
[16]
Saito , Mixed Hodge modules , Publ
M. Saito , Mixed Hodge modules , Publ. RIMS, Kyoto Univ. 26 (1990)
1990
-
[17]
Soergel , -cohomology of simple highest weight modules on walls and purity , Invent
W. Soergel , -cohomology of simple highest weight modules on walls and purity , Invent. Math. 98 (1989)
1989
-
[18]
Springer , A purity result for fixed point varieties in flag manifolds , J
T.A. Springer , A purity result for fixed point varieties in flag manifolds , J. Fac. Sci. Univ. Tokyo Sect. IA Math. 31 (1984)
1984
-
[19]
Verdier Sp\'ecialisation de faisceaux et monodromie mod\'er\'ee , Ast\'erisque 101 (1983)
J.L. Verdier Sp\'ecialisation de faisceaux et monodromie mod\'er\'ee , Ast\'erisque 101 (1983)
1983
-
[20]
Williamson , Nearby cycles without a function , https://mathoverflow.net/q/389689
G. Williamson , Nearby cycles without a function , https://mathoverflow.net/q/389689
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.