{"id":"84daedf7-dc50-4653-9df6-758e5975d843","arxiv_id":"2504.16365","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Nearby and vanishing cycle functors on rigid analytic varieties preserve Zariski-constructibility, satisfy Beilinson gluing, are perverse t-exact, and commute with Verdier duality.","lead":"This paper develops nearby and vanishing cycles for Zariski-constructible sheaves on rigid analytic varieties, extending the classical theory to the p-adic setting. It proves constructibility preservation, a Milnor fibre interpretation, Beilinson gluing, perverse t-exactness, and Verdier duality, laying groundwork for p-adic microlocal geometry.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.1(4) claims commutation with Verdier duality, but the proof of h is admitted (footnote 2, Remark 5.9(2)) to be noncanonical and incomplete; this directly weakens the central claim unless resolved.","rationale":"The reader identified imported resolution-of-singularities and Abhyankar results as the weakest premise. That is a legitimate concern, but those are standard theorems and the cited numbers are plausibly correct; the paper's own footnotes point to a more direct soft spot in the main duality theorem. Footnote 2 and Remark 5.9(2) explicitly disclaim canonicity and compatibility of α with the usual isomorphism, which conflicts with the abstract's promise that vanishing cycles \"commute with Verdier duality\" and with the theorem's use of the canonical-isomorphism symbol ≃. The proof of h also leaves several homotopy verifications and the general dévissage for α implicit. This is not a fatal error but an unresolved gap in the central claim, so the appropriate verdict is conditional acceptance pending either a proof of canonicity or a weakened statement using ≈ in place of ≃. The reader's worry about Temkin and Abhyankar should still be checked, but it is secondary in my assessment because those imported theorems are likely valid and are cited with precise references.","tokens_in":28378,"tokens_out":23086,"duration_ms":252413,"concrete_test":"Recompute the construction of Theorem 5.8(2) on the model case X=A^1, f=id, F=Λ (and also F=j_!Λ on G_m), taking two different topological generators t and t' of μ. Compute the induced h from Diagrams 5.8.1–5.8.7 and compare it with the classical canonical isomorphism obtained by dualizing the can/var triangle. If the two generators give different h, or if h differs from the classical isomorphism, then the canonically stated Theorem 1.1(4) is false as written and the theorem must be rephrased with a noncanonical ≈; if they agree, the remaining issue is to prove canonicity for general F, which would still be needed before accepting the claim at face value.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central duality assertion is Theorem 1.1(4) (Theorem 5.8(2)), which is stated with the paper's symbol for canonical isomorphism, ≃, and is highlighted in the abstract as \"commute with Verdier duality.\" Yet footnote 2 says \"We leave it to the reader to decide if it is canonical or not,\" and Remark 5.9(2) says the authors do not know whether the map α in Diagram 5.8.1 coincides with the usual isomorphism. The proof of h is a long chain of homotopy computations with several \"as one may verify\" steps and an unproved dévissage that α is an isomorphism for general F (only F supported on X0 or F≃j_*j^*F are explicitly discussed). If h is not canonical, or if it is not compatible with the usual duality for the can and var triangles, the paper proves only the existence of some natural isomorphism in the derived category, not the canonical commutation with Verdier duality promised in the abstract. This is a load-bearing caveat located inside the main theorem, not in an imported background result.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":28564,"tokens_out":7566,"duration_ms":73483,"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":[{"comment":"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.","section":"Theorem 5.8(2), Step 3; Remark 5.9(2)"},{"comment":"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.","section":"Theorem 5.1(2), proof of C(F)=0"}],"minor_comments":[{"comment":"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":"Theorem 1.1 and abstract"},{"comment":"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.","section":"Section 2, notation"},{"comment":"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.","section":"Lemma 2.12"},{"comment":"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.","section":"Proposition 4.19"},{"comment":"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.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious and potentially important contribution, but I cannot recommend acceptance while the central duality theorem for vanishing cycles has an admitted gap. The reader's report recommending acceptance appears to underestimate this issue. The authors should be asked to complete the proof of Theorem 5.8(2) for general F or to weaken the claims and abstract accordingly; the same applies to the delegated steps in the proof of Theorem 5.1(2). The reliance on Temkin's desingularization and Abhyankar's lemma is reasonable, as these are published results, but the manuscript should state the precise dependence explicitly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe 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.\n\nWhat 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.\n\nThe 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.\n\nWho 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.","headline":"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.","tokens_in":29124,"tokens_out":3180,"would_cite":true,"duration_ms":32264,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F20","14G22"],"pacs":[],"model":"deepseek-v4-flash","headline":"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.","keywords":["nearby cycles","vanishing cycles","rigid analytic varieties","Zariski-constructible sheaves","perverse sheaves","Verdier duality","Beilinson gluing","Milnor fibre"],"falsifier":"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.","tokens_in":28128,"feed_emoji":"","tokens_out":10175,"duration_ms":86644,"temperature":0.7,"pith_summary":"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.","feed_headline":"Rigid nearby cycles keep perverse and duality structure","feed_subtitle":"Finite-coefficient sheaves on non-archimedean spaces now have a full monodromy toolkit.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Builds the six-functor formalism and perverse sheaf theory for Zariski-constructible sheaves that the paper uses as its base context.","marker":"[BH22]"},{"why":"Supplies the resolution-of-singularities input that reduces finiteness and duality to strictly monomial divisors.","marker":"[Tem18]"},{"why":"Provides the rigid Abhyankar lemma used to trivialize local systems after Kummer base change in the duality proof.","marker":"[LP19]"},{"why":"Gives the finiteness and overconvergent-quasi-constructibility results used for the Milnor-fibre interpretation.","marker":"[Hub98]"},{"why":"Supplies the Iwasawa twist formalism and the duality and base-change lemmas used for the canonical monodromy statements.","marker":"[LZ19]"},{"why":"Carries Beilinson's construction of unipotent nearby cycles and maximal extension, which Section 4 adapts.","marker":"[Mor18]"},{"why":"Provides the classical construction of the duality map between nearby cycles and its reduction to the monomial case.","marker":"[Ill94]"},{"why":"Gives the gluing theorem for perverse sheaves that the paper translates to rigid analytic varieties.","marker":"[Bei87]"}],"fun_headline_variants":["Rigid nearby cycles: perverse t-exact and dual-friendly","Vanishing cycles on rigid spaces get Milnor fibres and gluing","Rigid vanishing cycles: perverse, dual, and constructible","Beilinson gluing and Milnor fibres for rigid vanishing cycles","Non-archimedean nearby cycles: perverse t-exact and dual"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Rigid nearby cycles: perverse t-exact and dual-friendly","Vanishing cycles on rigid spaces get Milnor fibres and gluing","Rigid vanishing cycles: perverse, dual, and constructible","Beilinson gluing and Milnor fibres for rigid vanishing cycles","Non-archimedean nearby cycles: perverse t-exact and dual"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000214,"raw_usage":{"total_tokens":1376,"prompt_tokens":847,"completion_tokens":529,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":463,"completion_tokens_details":{"reasoning_tokens":433}},"tokens_in":463,"tokens_out":529,"duration_ms":5443,"temperature":1.0,"reasoning_tokens":433,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:05:56.872625+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}