{"id":"aaa5f534-1707-4bbf-9e08-4d7c1814532f","arxiv_id":"2505.24599","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper constructs a universal coefficient category for wild Betti sheaves with an exponential local system and a Fourier equivalence, and identifies it as a canonical nontrivial R>0-torsor over Spec(Z).","lead":"This paper builds a universal category of \"wild sheaves\" that supports exponential functions and a Fourier transform on all spaces. A reader may care because the result uncovers a surprising hidden symmetry over the integers that connects algebraic geometry, symplectic geometry, and number theory.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The universality of W and the R_{>0}-torsor both rest on an unproved monoidal form of Vaintrob's Theorem 2 and an unchecked decomposition into 0- and ∞-factors; as written, Sections 5–6 are conditional on that input.","rationale":"The reader's weakest_assumption identifies the same load-bearing point: Proposition 5.1 is cited rather than proved, and the later universality and torsor statements are built on it. I agree with that assessment and do not think the verdict should move. I would add that the proof of universality needs not only the bare equivalence but also the monoidal refinement and the orthogonality of the two quotient factors; the paper's sketch does not supply these. Section 6 inherits the same gap. None of this casts doubt on the self-contained parts: the construction of W, the exponential local system exp, the vanishing π_!exp=0, and the Fourier equivalence of Theorem 4.1 are all argued in enough detail to be checked. The paper would become acceptable as a research announcement once Vaintrob's theorem, or a direct proof, is available in the needed monoidal form and the quotient decomposition is written out. Since the reader already marked the paper CONDITIONAL for essentially this reason, the correct adjustment is to leave the verdict unchanged.","tokens_in":7984,"tokens_out":22445,"duration_ms":294655,"concrete_test":"Give a complete monoidal proof of Proposition 5.1. In particular, on the generators corresponding to skyscraper sheaves δ_r, check that the tensor product of their images is the image of δ_r⋆δ_s=δ_{r+s}; if the monoidal structure is only checked on underlying objects, the claimed symmetric monoidal equivalence is not established. Then prove that the quotient by the point decomposes into two orthogonal factors, verifying on the same generators that Hom between the 0- and ∞-factors vanishes, and that the 0-factor is W. This directly tests the implication used in Section 5 and inherited by Proposition 6.1.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The pivot of the paper is Proposition 5.1, a stated variant of Theorem 2 of Vaintrob's unpublished notes. It is not proved: the proof given is a sketch that glues two chart equivalences. Section 5 then needs more than a bare equivalence of categories. It needs the equivalence to be symmetric monoidal, and it needs the quotient Dqc(˜P1,a/˜Gm)/Dqc(˜Gm/˜Gm) to decompose as a product of two orthogonal factors, one equivalent to W and the other to W with reversed filtration. The text only asserts this decomposition. If the two factors form only a recollement or have nonzero Homs between them, then a coefficient category can be linear over gluing data rather than over W, and the statement that W is the universal category of coefficients has no clear content. The same unproved decomposition is reused in Proposition 6.1 to identify W ⊗_S W with D(R>0,W), so the claimed canonical nontrivial R>0-torsor over Spec(S) inherits the dependency. This is not an internal inconsistency in the main construction: the definition of W, the exponential local system, and the Fourier equivalence in Sections 2–4 are self-contained and convincing. The weak point is precisely the bridge from the free convolutive category to the almost algebraic stack, and from that stack to a product decomposition.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs an enlargement of Betti sheaves by working with coefficients in W, the symmetric monoidal stable infinity-category of completely and continuously R-filtered spectra (Definition 2.2). It defines wild sheaves D(X,W), constructs an invertible exponential local system exp on R whose fibres are the shifted units S(r) (Proposition 3.1), proves the key vanishing pi_! exp = 0 (Proposition 3.3), and establishes a Fourier equivalence for all real vector spaces (Theorem 4.1). The paper then claims in Section 5 that W is universal among coefficient categories admitting a nontrivial exponential local system inducing a Fourier equivalence, using Proposition 5.1, a variant of an unpublished theorem of Vaintrob. Section 6 claims that Spec(W) over Spec(S) is a canonical nontrivial R_{>0}-torsor, with Proposition 6.1 identifying W tensor_S W with D(R_{>0},W).","tokens_in":8289,"tokens_out":3296,"duration_ms":41484,"significance":"The construction in Sections 2–4 is explicit, self-contained, and convincing: the definition of W, the exponential local system, the vanishing of pi_! exp, and the Fourier equivalence are all proved directly from the definitions. If the universality and torsor claims hold, the paper would give a clean conceptual explanation of Tamarkin's enhanced sheaves and a genuinely surprising new structure over Spec(Z). The elementary parts of the paper are a solid contribution. However, the central novelty is not yet established at the level of rigor required for a journal publication, because it depends on Proposition 5.1 and on a decomposition statement that are only sketched and partly rely on an unpublished source.","major_comments":[{"comment":"Proposition 5.1 is the pivot of the universality argument, but it is stated as a symmetric monoidal equivalence and the proof given is only a sketch, with the citation referring to an unpublished note [Vai17]. The monoidal structure is essential: the universality statement concerns coefficient categories with a convolution structure, so the equivalence must preserve the symmetric monoidal structure, not merely the underlying category. This needs either a complete proof in the paper or a precise reference to a published or otherwise fully available source.","section":"§5, Proposition 5.1"},{"comment":"The assertion that the quotient Dqc(˜P1,a/˜Gm)/Dqc(˜Gm/˜Gm) decomposes as a product of two orthogonal factors, one equivalent to W and the other to W with reversed filtration, is load-bearing but unproved. If the two factors are only related by a recollement or have nontrivial Homs, a coefficient category could be linear over gluing data rather than over W, and the claimed universality would not follow. The paper should either prove the decomposition, including the claimed equivalence of each factor, or identify a published reference that contains it.","section":"§5, paragraph after Proposition 5.1"},{"comment":"The proof of Proposition 6.1, and hence the entire torsor statement, inherits the same unproved decomposition: after base change to W, the text asserts that 'removing this part, both sides decompose into two pieces' and concludes W ⊗_S W ≅ D(R_{>0},W). Since the paper notes that 'this equivalence can be proved directly,' such a direct proof should be supplied, or the missing decomposition should be established. As written, the torsor claim is conditional on the same external input as the universality claim.","section":"§6, Proposition 6.1"}],"minor_comments":[{"comment":"The word 'introcued' in the first sentence of Section 2 should be 'introduced'.","section":"§2, Definition 2.2"},{"comment":"The word 'paranthetical' should be 'parenthetical'.","section":"§6, Warning 6.2"},{"comment":"The phrase 'cf. below' in the sketch of Proposition 5.1 is vague; the relevant equivalence could be stated explicitly in the sketch.","section":"§5, first paragraph"},{"comment":"The notation R_{>0,Betti} := Spec(D(R_{>0},S)) is used before the general categorical notion of Spec is introduced; a brief explanation of this notation would improve readability.","section":"§6, notation"},{"comment":"Reference [Vai17] is unpublished and appears only as a URL; if it remains unpublished, the paper should include a more precise statement of the theorem being used or an appendix proving the needed variant.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is written by a leading author and the elementary parts are clean, but the main claims in Sections 5–6 rest on a sketch of an unpublished theorem and an unproved decomposition. This is a conditional situation rather than an internal inconsistency, so major revision seems appropriate: the author should either supply the missing proofs or point to a published reference containing them. The note format should not exempt a central proposition from a complete proof."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear X,\n\nYou should read this as two papers. Sections 2–4 give a clean, self-contained construction of the coefficient category W and the exponential local system, and the Fourier equivalence is proved from the definitions. That part is solid and worth having on record even though much of it recovers Tamarkin's enhanced sheaves, which Scholze credits explicitly. The vanishing pi!exp=0, the kernel computation for the Fourier transform, and the Legendre transform for convex functions are all done properly.\n\nThe novel part is Sections 5–6: the universality of W and the claim that Spec(W) is a nontrivial R>0-torsor over Spec(Z). These are genuinely new, but they rest on Proposition 5.1, a 'variant' of an unpublished theorem of Vaintrob, and the proof given is a sketch. The stress-test note is right about what is missing: you need more than an equivalence of categories. You need it to be symmetric monoidal, and you need the quotient Dqc(~P1,a/~Gm)/Dqc(~Gm/~Gm) to decompose as a product of two orthogonal factors, one W and one W with reversed filtration. The text asserts this decomposition without proof. If it is only a recollement, or if there are Homs between the two factors, then the universality statement overreaches. The torsor statement in Section 6 inherits the same dependency. This is not an internal contradiction; it is a load-bearing external premise.\n\nI would not call the paper misleading. It is clearly a note, it credits Tamarkin and Vaintrob, and it flags the sketchiness implicitly. But as written, the headline claims go beyond what is demonstrated. A referee should be able to verify the Vaintrob input or ask for the full proof of Proposition 5.1 and the decomposition before signing off.\n\nWho should read this? Anyone working in sheaf theory, symplectic geometry, or exponential motives. The W construction and the exp local system are useful tools even if the universality part is provisional. I'd bring it to reading group and would cite the clean construction.\n\nRecommendation: send to refereeing, not desk reject, but the referee should focus on Proposition 5.1 and the decomposition. If those are not supplied, the paper should be revised to present Sections 5–6 as conditional.","headline":"Solid construction of wild sheaves, but the universality and torsor claims are conditional on an unproved variant of Vaintrob's theorem and a sketchy decomposition.","tokens_in":8828,"tokens_out":1648,"would_cite":true,"duration_ms":19457,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F08","14F05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Wild Betti sheaves are universal, revealing a hidden torsor","keywords":["wild Betti sheaves","exponential local system","Fourier transform","enhanced sheaves","R-filtered spectra","R>0-torsor","Betti sheaves","six operations"],"falsifier":"Exhibit a presentable symmetric monoidal stable $\\infty$-category $C$ that admits a nontrivial additive exponential local system inducing a Fourier equivalence but whose associated symmetric monoidal functor from the almost infinite-root stack to $C$ does not factor through $W$ (up to reversing the filtration); this would disprove universality without challenging the existence theorem. Alternatively, complete the proof of the cited stack equivalence and check directly whether the two quotient categories at $0$ and $\\infty$ are exactly $W$ and its filtration-reverse.","tokens_in":7777,"feed_emoji":"🔁","tokens_out":13190,"duration_ms":151314,"temperature":0.7,"pith_summary":"The paper builds a universal enlargement of the category of Betti sheaves, called wild Betti sheaves, that carries a nontrivial \"exponential local system\" on the real line and a Fourier equivalence defined on every sheaf. The universal coefficient category is the category $W$ of completely and continuously $\\mathbb{R}$-filtered spectra, and the construction recovers a previously known theory of enhanced sheaves. The main structural conclusion is that $\\mathrm{Spec}(W)$ sits over $\\mathrm{Spec}(\\mathbb{Z})$ as a canonical nontrivial $\\mathbb{R}_{>0}$-torsor, which gives a new direction in which to localize in algebraic and arithmetic geometry. This matters because a Fourier transform is the engine behind many applications to exponential sums and geometric representation theory, and the paper shows it is available in the Betti setting in a way that does not depend on arbitrary choices.","feed_headline":"Wild Betti sheaves are universal, revealing a hidden torsor","feed_subtitle":"The universal coefficient category, W, is a canonical nontrivial R>0-torsor over Spec(Z).","key_machinery":"The central object is $W$, the symmetric monoidal stable $\\infty$-category of completely and continuously $\\mathbb{R}$-filtered spectra: an object is a spectrum with an ascending filtration $\\mathrm{Fil}_r M$ indexed by $r\\in\\mathbb{R}$, complete at $-\\infty$ and continuous at every $r$. Because the unit $S(0)$ is not compact, this category escapes the usual obstruction that makes all invertible sheaves on $\\mathbb{R}$ trivial; the exponential local system is the family of shifts $S(f)$ with filtration jumping at $f(x)$. The argument is carried by the kernel $K_V=b^*\\exp$ on $V\\times V^*$, whose convolution gives the Fourier equivalence, and by a cited equivalence between sheaves on $\\mathbb{R}$ with the convolution product and quasicoherent sheaves on an almost infinite-root version of $\\mathbb{P}^1/\\mathbb{G}_m$, which allows the universal property to be read off from the two ends $0$ and $\\infty$ of that stack.","core_discovery":"The central discovery is that the problem of adjoining an exponential local system to Betti sheaves has a universal solution. Concretely, the paper proves existence of a coefficient category $W$ whose objects are completely and continuously $\\mathbb{R}$-filtered spectra, together with an invertible sheaf $\\exp\\in D(\\mathbb{R},W)$ whose fibre at $r$ is the shift $S(r)$; from this kernel one defines a Fourier functor $F_V$ on every real vector space $V$, and Theorem 4.1 says $F_V$ is an equivalence with inverse $(-1)^*F_{V^*}[d]$. The universality statement of Section 5 says that, up to reversing the filtration, $W$ is the unique such coefficient category, because any candidate exponential local system inducing a Fourier equivalence is forced to be trivial after restriction to $\\tilde{\\mathbb G}_m/\\tilde{\\mathbb G}_m$ and the remaining quotient splits into $W$ at $0$ and its filtration-reverse at $\\infty$. The final structural consequence is that $\\mathrm{Spec}(W)\\to\\mathrm{Spec}(\\mathbb{S})$ is a torsor under $\\mathbb{R}_{>0,\\mathrm{Betti}}=\\mathrm{Spec}(D(\\mathbb{R}_{>0},\\mathbb{S}))$, hence a canonical nontrivial $\\mathbb{R}_{>0}$-torsor over $\\mathrm{Spec}(\\mathbb{Z})$.","pith_inferences":["If the ring-stack formulation from the outlook is taken literally, each realization of sheaf theory should produce its own canonical torsor; comparing the Betti torsor with the torsors attached to other realizations would constrain any lifting of exponential motives to a global category.","The nonconvex case, where the Fourier transform of $S(f)$ is noninvertible and its fibre at $0$ records contributions from local extrema, suggests using wild Betti sheaves as a microlocal bookkeeping device for oscillatory integrals, a use the paper does not develop.","A direct verification that the coaction sends $(\\mathrm{Fil}_r M)_r$ to the sheaf with stalk $\\mathrm{Fil}_{rt}M$ at $t$ would give an independent check of the torsor structure without passing through the cited stack equivalence."],"forward_implications":["Wild sheaves on any space $X$ form a six-operation category $D(X,W)$ containing ordinary sheaves fully faithfully, so existing sheaf-theoretic arguments transfer to the enlarged setting.","For every continuous $f:X\\to\\mathbb{R}$ there is a canonical invertible wild sheaf $S(f)$, making any variety with potential $(X,f)$ into a wild Betti sheaf whose cohomology is a candidate for exponential or rapid-decay cohomology.","The Fourier transform is an equivalence on all sheaves, intertwines convolution with tensor product, and for convex functions on $\\mathbb{R}$ sends $S(f)$ to $S(f^*)[-1]$, where $f^*$ is the Legendre transform.","If the universality claim is right, the whole theory lives over a canonical nontrivial $\\mathbb{R}_{>0}$-torsor over $\\mathrm{Spec}(\\mathbb{Z})$, and after base change wild sheaves on $X$ become a twisted form of sheaves on $X\\times\\mathbb{R}_{>0}$.","The same construction applies with coefficients in complete almost modules over the Novikov ring or over any rank-one valuation ring, matching the motivic examples from almost mathematics."],"supporting_citations":[{"why":"Supplies the equivalence between sheaves on $\\mathbb{R}$ with convolution and quasicoherent sheaves on an almost infinite-root version of $\\mathbb{P}^1/\\mathbb{G}_m$; the paper gives only a sketch of this proposition.","marker":"[Vai17]"},{"why":"Constructs the enhanced-sheaf category and proves the Fourier equivalence that Theorem 4.1 extends.","marker":"[Tam18]"},{"why":"Identifies the enhanced-sheaf construction in the irregular Riemann-Hilbert correspondence.","marker":"[DK16]"},{"why":"Provides the formalism used to extend the six operations to wild sheaves on locally compact Hausdorff spaces.","marker":"[HM24]"},{"why":"Discusses equivalent incarnations of the enhanced-sheaf category, used to connect $W$ with existing formulations.","marker":"[KZ25]"}],"fun_headline_variants":["Wild Betti sheaves unmask a hidden R>0-torsor","Universal Betti sheaf category is a nontrivial R>0-torsor over Z","Canonical nontrivial R>0-torsor emerges from wild Betti sheaves","Unique wild Betti sheaf category forces a nontrivial torsor over Z","Wild Betti sheaves: universal Fourier equivalence reveals hidden torsor"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of universality depends on an imported equivalence between sheaves on $\\mathbb{R}$ with convolution and quasicoherent sheaves on an almost infinite-root version of $\\mathbb{P}^1/\\mathbb{G}_m$, which is only sketched and cited to unpublished work; if that equivalence or the claimed splitting at $0$ and $\\infty$ fails, the universality of $W$ and the torsor statement would collapse.","fun_headline_variants_meta":{"raw":{"variants":["Wild Betti sheaves unmask a hidden R>0-torsor","Universal Betti sheaf category is a nontrivial R>0-torsor over Z","Canonical nontrivial R>0-torsor emerges from wild Betti sheaves","Unique wild Betti sheaf category forces a nontrivial torsor over Z","Wild Betti sheaves: universal Fourier equivalence reveals hidden torsor"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002239,"raw_usage":{"total_tokens":8642,"prompt_tokens":916,"completion_tokens":7726,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":532,"completion_tokens_details":{"reasoning_tokens":7618}},"tokens_in":532,"tokens_out":7726,"duration_ms":61151,"temperature":1.0,"reasoning_tokens":7618,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:18:35.512559+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a presentable symmetric monoidal stable $\\infty$-category $C$ that admits a nontrivial additive exponential local system inducing a Fourier equivalence but whose associated symmetric monoidal functor from the almost infinite-root stack to $C$ does not factor through $W$ (up to reversing the filtration); this would disprove universality without challenging the existence theorem. Alternatively, complete the proof of the cited stack equivalence and check directly whether the two quotient categories at $0$ and $\\infty$ are exactly $W$ and its filtration-reverse.","supporting_citations":[],"review_version":1}