Vanishing Cycles for Zariski-Constructible Sheaves on Rigid Analytic Varieties

Tong Zhou

REVIEW 2 major objections 5 minor 5 references

Vanishing Cycles for Zariski-Constructible Sheaves on Rigid Analytic Varieties

T0 review · 2 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper establishes that nearby and vanishing cycles for Zariski-constructible sheaves on rigid analytic varieties preserve constructibility, are perverse t-exact, and satisfy Verdier duality up to Tate and Iwasawa twists.

desk verdict Solid development of nearby/vanishing cycles for Zariski-constructible sheaves on rigid spaces, but the central Verdier-duality claim for vanishing cycles is not proved as fully as advertised. read the letter →

arxiv 2504.16365 v1 pith:RE3QL6JJ submitted 2025-04-23 math.AG

Tong Zhou This is my paper · ORCID

classification math.AG MSC 14F2014G22

keywords nearbycyclesvanishingrigidanalyticvarietiesZariski-constructiblesheavesperverseVerdierdualityBeilinsongluingMilnorfibre

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

Over a complete, algebraically closed non-archimedean field, the paper defines nearby and vanishing cycle functors for finite-coefficient Zariski-constructible sheaves on rigid analytic varieties, in families $X \to \mathbb{A}^1$. It proves that these functors preserve Zariski-constructibility, admit a Milnor-fibre description of their stalks, satisfy Beilinson's gluing construction, are perverse t-exact, and commute with Verdier duality up to the expected Tate and Iwasawa twists. The outcome is that the classical monodromy toolkit of complex and algebraic geometry, which studies how cohomology changes as a parameter approaches a special value, works in the rigid analytic setting as well. This matters because the resulting functors are the basic input for a microlocal sheaf theory on rigid spaces, an application flagged in the paper.

What carries the argument

The load-bearing object is the nearby-cycle functor $\psi_f(F) = \varinjlim_n i^* j_* p_{n*} p_n^* F$, built from the Kummer coverings $p_n \colon \mathbb{G}_m \to \mathbb{G}_m$ of the punctured line; $\varphi_f(F)$ is its cone over the specialization map. The Iwasawa twist $(-1)^\tau$ removes the choice of a topological generator of the monodromy group, which is what makes the duality statements canonical. The reduction mechanism is a two-step dévissage: first resolution of singularities makes the special fibre strictly monomial, then étale coordinates and Abhyankar's lemma reduce the duality check to an algebraic comparison. The key triangle $i^* j_* j^* F \to \psi_f(F) \to \psi_f(F)(-1)^\tau \to$ organizes the monodromy action and drives the duality diagram.

What would settle it

Run the duality comparison on the strictly monomial example $X = \operatorname{Spa}(K\langle T_1,\dots,T_n\rangle)$ with $f = T_1^{n_1}\cdots T_r^{n_r}$ and $F = j_! L$ for a local system $L$ with non-trivial monodromy: if $\Psi_f D F \to (D\Psi_f F)(1)$ is not an isomorphism on stalks at a point of the special fibre, the theorem is false; alternatively, search for a singular $X$ where $\psi_f(F)$ fails to be Zariski-constructible.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: for $\Lambda = \mathbb{F}_{\ell^r}$ or $\mathbb{Z}/\ell^r$ with $\ell \neq p$, the shifted functors $\Psi_f = \psi_f[-1]$ and $\Phi_f = \varphi_f[-1]$ attached to a map $f \colon X \to \mathbb{A}^1$ are perverse t-exact in the specified senses. There are canonical isomorphisms $\Psi_f D F \simeq (D \Psi_f F)(1)$ and $\Phi_f D F \simeq (D \Phi_f F)(-1)^\tau (1)$, and the specialization, canonical, and variation triangles are dual to each other via Verdier duality. The paper also proves constructibility preservation, a Milnor-fibre interpretation of stalks, Beilinson's gluing equivalence, and compatibility with smooth pullbacks and quasi-compact quasi-separated pushforwards. The argument proceeds by a dévissage to strictly monomial divisors through resolution of singularities, followed by comparison with the algebraic case.

Load-bearing premise

The argument depends on Temkin resolution of singularities and Abhyankar's lemma for rigid analytic varieties over algebraically closed non-archimedean fields; if either fails in the needed generality, the constructibility, perverse t-exactness, and duality conclusions can collapse.

Editorial extensions

If this is right

Zariski-constructible sheaves on rigid analytic varieties now have a monodromy theory: nearby and vanishing cycles land in the same constructible derived category.
Beilinson's gluing theorem holds in this context, so perverse sheaves on the total space are equivalent to gluing data on the punctured part and the special fibre.
The Milnor-fibre description gives a concrete way to compute stalks of vanishing cycles as tubes around the special fibre.
Verdier duality for nearby and vanishing cycles with Tate and Iwasawa twists means the standard six-functor arguments about cycles, including duals of specialization and variation maps, work rigid-analytically.

Reading between the lines

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

A natural next test is a Thom-Sebastiani or Künneth formula for these functors, which the paper lists as an open question; if it holds, vanishing cycles on product families would decompose as external tensor products.
The comparison between the paper's $\varphi$-ULA condition and the existing notion of universal local acyclicity would decide whether these cycles can serve as a local acyclicity criterion in arithmetic geometry.
If the tor-finite restriction in the $\mathbb{Z}/\ell^r$ statement is essential, dropping it may produce counterexamples involving infinite-rank stalks; the direct-sum example in Remark 3.2 shows extension to non-extendable sheaves already fails.

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

Dear colleague,

The two things you should know: this preprint builds the expected nearby/vanishing cycle formalism for Zariski-constructible sheaves on rigid analytic varieties, and most of it works. The catch is in the final theorem: the asserted commutation of vanishing cycles with Verdier duality is not proved in the strong form the abstract suggests, and the authors essentially concede this in a footnote.

What is genuinely new and good: the author defines psi_f and phi_f with finite coefficients, using the Iwasawa twist to avoid choosing a topological generator, and proves the expected basic properties (smooth base change, proper pushforward, analytification). Proposition 3.1 gives constructibility preservation, with a Milnor fibre/tube interpretation in Proposition 3.8. Section 4 adapts Beilinson's construction and gluing; Section 5 proves perverse t-exactness of Psi, and of Phi for tor-finite coefficients. The adaptation of Mor18, Bei87, and Ill94 is competent, and the paper is careful about where non-quasi-compactness of X^× forces changes to the usual arguments. The paper is also honest about open questions and its own uncertainties.

The soft spot is Theorem 5.8(2). The construction of h is a long chain of homotopy computations, several steps left as "one may verify," and the proof that the map alpha is an isomorphism only treats two cases explicitly (F supported on X0 and F isomorphic to j_*j^*F); the general case is asserted without the missing dévissage. Remark 5.9(2) states that the authors do not know whether alpha is the usual isomorphism. So what is proved is the existence of some natural isomorphism h, not the canonical commutation with Verdier duality promised in the abstract. This matters: the duality triangle (Diagram 5.8.8) is a headline result. I suspect it is fixable, but it is not fixed here. The delegated arguments (Lemma 2.12, Proposition 4.19) are less worrying, though a referee should ask for the omitted diagram chase and the details in the Beilinson gluing proof. The reliance on Temkin's resolution of singularities and on Abhyankar's lemma in the rigid setting is imported background, used centrally; if those results were to fail in this generality the whole dévissage would break, but they are published results and the paper cites them honestly.

Who this is for: anyone working with constructible sheaves on rigid analytic spaces, especially those who need nearby cycles, perverse sheaves, or the microlocal theory the author says is coming. It deserves a serious referee; the referee's main job should be to push on Theorem 5.8(2) and the canonicity of h.

Referee Report

2 major / 5 minor

Summary. The paper develops a theory of nearby and vanishing cycles for Zariski-constructible étale sheaves with finite coefficients on rigid analytic varieties over an algebraically closed non-archimedean field of characteristic zero or mixed characteristic. It defines ψ_f and φ_f, proves preservation of Zariski-constructibility, gives a Milnor fibre description of stalks, adapts Beilinson's gluing construction, establishes perverse t-exactness, and states a Verdier duality result for both functors. The main technical devices are the Iwasawa twist to make the monodromy action canonical, reduction to strictly monomial divisors via Temkin's desingularization, and Abhyankar's lemma in the rigid setting.

Significance. If the main theorems hold, the paper fills a genuine gap by providing a full nearby/vanishing cycle formalism in the Zariski-constructible rigid analytic setting, with the expected finiteness, perversity, and duality properties. It will be a foundational input for the author's planned microlocal sheaf theory. The paper also contains useful preparatory results, including General Artin-Grothendieck Vanishing and a treatment of tor-finite coefficients. However, the advertised Verdier duality for vanishing cycles is not fully established as stated, and the proof of the canonical duality for nearby cycles contains several delegated steps; these are load-bearing issues for the central claims.

major comments (2)

[Theorem 5.8(2), Step 3; Remark 5.9(2)] The proof that the map α (and hence h) in Diagram 5.8.1 is an isomorphism treats only the special cases F ≃ j_*j^*F and F ≃ i_*i^*F, and no reduction to these cases is provided for a general F ∈ D^b_zc(X). Moreover, footnote 2 and Remark 5.9(2) explicitly state that the authors do not know whether h is canonical or whether α coincides with the usual duality isomorphism. Since the abstract and Theorem 1.1(4) claim that vanishing cycles 'commute with Verdier duality' and use the symbol ≃, the central statement of the paper is not established as written. The authors should either complete the dévissage for general F or revise the theorem and abstract to state precisely which assertions are proven.
[Theorem 5.1(2), proof of C(F)=0] The proof that the canonical map g for nearby cycles is an isomorphism relies on several unverified assertions: that the cone C(F) commutes with proper pushforwards and analytification, that a 'standard dévissage and induction on dim supp(F)' reduces to F = j_!F_U, and that after applying Abhyankar's lemma the situation becomes algebraisable so that the algebraic comparison applies. These steps are load-bearing because Theorem 1.1(3) is advertised as canonical. The passage from the local étale model to the algebraic counterpart should be written out or supplied with precise references.

minor comments (5)

[Theorem 1.1 and abstract] The abstract and Theorem 1.1(4) use 'commute with Verdier duality' and the symbol ≃, while the theorem itself only asserts a 'natural isomorphism' h and footnote 2 adds a caveat about canonicity. The statements should be made consistent.
[Section 2, notation] The topos X_0 ¯× B_μ is used before its definition in footnote 5; the authors should define this notation in the main text of Section 2.
[Lemma 2.12] The proof is summarized as a 'standard diagram chase' with no details. Since the lemma is used later (e.g., in Proposition 3.1), at least an indication of which base-change and colimit commutation results are involved would aid verification.
[Proposition 4.19] The authors state that 'the original proof in [Bei87, 3.1]' applies verbatim, but they do not spell out which properties of the rigid analytic setting ensure that the gluing construction works. A short discussion of the axioms being verified would make the adaptation transparent.
[Throughout] The phrase 'ℓ = p ≠ 0 allowed' in Section 2 appears to be a rendering of 'ℓ ≠ p'; the coefficient characteristic assumptions should be stated unambiguously. The final PDF should also be proofread, as the provided text contains several OCR artifacts.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the derivation is self-contained; cited external results are independent, and the admitted noncanonicity of h is a caveat, not a circular reduction.

full rationale

The paper's nearby and vanishing cycle functors are defined directly (Definition 2.1), and its finiteness, perversity, and duality assertions are proved from those definitions plus imported, independent results. The finiteness proof (Proposition 3.1) uses Temkin's resolution of singularities and Huber's finiteness theorems; the duality proof (Theorem 5.1(2), Theorem 5.8(2)) reduces, after resolution and Abhyankar's lemma, to the known algebraic comparison results in [Ill94] and [LZ19]. None of these imports is an instance of the theorem being proved, nor are the analytic functors defined in terms of their own duality or perversity behavior. The only self-references are forward-looking citations to the author's forthcoming [Zho] for future applications, which are not load-bearing. Footnote 2 and Remark 5.9(2) explicitly admit that the map h in Theorem 5.8(2) is not known to be canonical and that α may not coincide with the usual isomorphism; this weakens the strength of the claimed duality statement but does not mean the result is assumed as an input. No fitted parameter is renamed as a prediction, and no uniqueness or ansatz is imported from the authors' prior work. The proof chain therefore has no circular step; any concern about completeness or canonicity belongs to correctness risk, not circularity.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The central theorems depend on imported results that are standard in the field but not proved here: the six-functor formalism for etale sheaves on adic spaces, resolution of singularities, Huber's finiteness and cohomological dimension theorems, Abhyankar's lemma in the rigid context, and the Beilinson-Morel-Illusie constructions. There are no fitted parameters and no invented entities; the new functors are explicit constructions. The finite-coefficient and l not equal to p hypotheses are stated as setup in section 1.

assumptions (7)

standard math Six-functor formalism for etale sheaves on adic spaces, enhanced to infinity-categories
The paper works inside the framework of [Hub96] as enhanced by [Zav24, section 9]; adopted in the Conventions and used throughout.
standard math Finite-coefficient etale sheaves on rigid analytic varieties have no wild ramification
Remark 2.3(2) cites [Lut93] for the fact that finite coefficients eliminate wild ramification, justifying the use of the finitary fundamental group in the definition of nearby cycles.
standard math Temkin's resolution of singularities for rigid analytic varieties
Used in Proposition 3.1 and Theorem 5.1(2) to reduce to strictly monomial divisors; cited as [Tem18, 1.1.13(i), 1.1.11].
standard math Huber's finiteness theorem and cohomological dimension bounds
Theorem 3.4 ([Hub98, 2.1]) and the bound in [Hub96, 2.8.3] underlie the Milnor fibre interpretation and the boundedness arguments in Lemma 5.3.
standard math Abhyankar's lemma in the rigid analytic context
Used in Theorem 5.1(2) to split covers after Kummer base change; cited as [LP19, Proposition 2.1].
standard math Beilinson's gluing theorem and Morel's construction
Proposition 4.19 and the maximal extension functor Xi are built on [Bei87, 3.1] and [Mor18].
domain assumption Base field and coefficient hypotheses: K non-trivially valued, complete, algebraically closed, mixed or equal characteristic zero, and Lambda finite with l not equal to p for the main theorems
Stated in the introduction (section 1) as the ambient setup; the l not equal to p condition is required for perverse t-exactness and Verdier duality.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Vanishing Cycles for Zariski-Constructible Sheaves on Rigid Analytic Varieties." pith.science (2026). https://pith.science/paper/RE3QL6JJ

@misc{pith2026250416365,
  author       = {Pith},
  title        = {Pith review of: Vanishing Cycles for Zariski-Constructible Sheaves on Rigid Analytic Varieties},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RE3QL6JJ}},
  note         = {Machine review of arXiv:2504.16365}
}

read the original abstract

We develop a theory of nearby and vanishing cycles in the context of finite-coefficient Zariski-constructible sheaves over a non-archimedean field which is non-trivially valued, complete, algebraically closed, and of mixed characteristic or equal characteristic zero. Apart from basic properties, we show that they preserve Zariski-constructibility, have a Milnor fibre interpretation, satisfy Beilinson's gluing construction, are perverse t-exact, and commute with Verdier duality.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

5 extracted references · 2 canonical work pages

[1]

Faisceaux pervers

[BBDG] Alexander Beilinson, Joseph Bernstein, Pierre Deli gne and Ofer Gabber. “Faisceaux pervers”. In: Astérisque 100 (2008) (↑ 9, 10). [Bei87] Alexander Beilinson. “How to glue perverse sheaves ”. In: K-Theory, Arithmetic and Geo- metry: Seminar, Moscow University, 1984–1986 . Ed. by Yuri I. Manin. Berlin, Heidelberg: Springer, 1987, pp. 42–51 ( ↑1, 12)...

arXiv 2008
[30]

A ﬁniteness result for direct image s heaves on the étale site of rigid analytic varieties

Aspects of Mathematics. Wiesbaden: Vieweg+Teubner Verlag , 1996 (↑2, 6, 7, 14). [Hub98] Roland Huber. “A ﬁniteness result for direct image s heaves on the étale site of rigid analytic varieties”. In: Journal of Algebraic Geometry 7.2 (1998), pp. 359–403 ( ↑8, 9). [Ill17] Luc Illusie. “Around the Thom–Sebastiani theorem, with an appendix by Weizhe Zheng”. ...

work page 1998
[223]

Vanishing cycle sheaves and hol onomic systems of diﬀerential equa- tions

talk:1. Société mathématique de France, 1994, pp. 9–57 ( ↑13). [Kas83] Masaki Kashiwara. “Vanishing cycle sheaves and hol onomic systems of diﬀerential equa- tions”. In: Algebraic Geometry. Ed. by Michel Raynaud and Tetsuji Shioda. Berlin, Heidel- berg: Springer Berlin Heidelberg, 1983, pp. 134–142 ( ↑1). [KS90] Masaki Kashiwara and Pierre Schapira. Sheav...

arXiv 2019
[305]

Springer

Lecture Notes in Mathematics. Springer. Dirigé par Michael Artin, Alexandre Grothendieck, and Jean-Louis Verdier, 1972, 1973 ( ↑5). [SGA 7] SGA

work page 1972
[340]

Functorial desingularization ov er Q: boundaries and the embedded case

Lecture Notes in Mathematics. Berlin, Heidelberg: Springer. I dirigé par Al exandre Grothendieck, II dirigé par Pierre Deligne and Nicholas Katz ( ↑1). [Stacks] The Stacks Project, https://stacks.math.columbia.edu/ . 2024 (↑15). [Tem18] Michael Temkin. “Functorial desingularization ov er Q: boundaries and the embedded case”. In: Israel Journal of Mathemat...

arXiv 2018

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

Faisceaux pervers

[BBDG] Alexander Beilinson, Joseph Bernstein, Pierre Deli gne and Ofer Gabber. “Faisceaux pervers”. In: Astérisque 100 (2008) (↑ 9, 10). [Bei87] Alexander Beilinson. “How to glue perverse sheaves ”. In: K-Theory, Arithmetic and Geo- metry: Seminar, Moscow University, 1984–1986 . Ed. by Yuri I. Manin. Berlin, Heidelberg: Springer, 1987, pp. 42–51 ( ↑1, 12)...

arXiv 2008

[2] [30]

A ﬁniteness result for direct image s heaves on the étale site of rigid analytic varieties

Aspects of Mathematics. Wiesbaden: Vieweg+Teubner Verlag , 1996 (↑2, 6, 7, 14). [Hub98] Roland Huber. “A ﬁniteness result for direct image s heaves on the étale site of rigid analytic varieties”. In: Journal of Algebraic Geometry 7.2 (1998), pp. 359–403 ( ↑8, 9). [Ill17] Luc Illusie. “Around the Thom–Sebastiani theorem, with an appendix by Weizhe Zheng”. ...

work page 1998

[3] [223]

Vanishing cycle sheaves and hol onomic systems of diﬀerential equa- tions

talk:1. Société mathématique de France, 1994, pp. 9–57 ( ↑13). [Kas83] Masaki Kashiwara. “Vanishing cycle sheaves and hol onomic systems of diﬀerential equa- tions”. In: Algebraic Geometry. Ed. by Michel Raynaud and Tetsuji Shioda. Berlin, Heidel- berg: Springer Berlin Heidelberg, 1983, pp. 134–142 ( ↑1). [KS90] Masaki Kashiwara and Pierre Schapira. Sheav...

arXiv 2019

[4] [305]

Springer

Lecture Notes in Mathematics. Springer. Dirigé par Michael Artin, Alexandre Grothendieck, and Jean-Louis Verdier, 1972, 1973 ( ↑5). [SGA 7] SGA

work page 1972

[5] [340]

Functorial desingularization ov er Q: boundaries and the embedded case

Lecture Notes in Mathematics. Berlin, Heidelberg: Springer. I dirigé par Al exandre Grothendieck, II dirigé par Pierre Deligne and Nicholas Katz ( ↑1). [Stacks] The Stacks Project, https://stacks.math.columbia.edu/ . 2024 (↑15). [Tem18] Michael Temkin. “Functorial desingularization ov er Q: boundaries and the embedded case”. In: Israel Journal of Mathemat...

arXiv 2018