REVIEW 4 major objections 5 minor 12 references
Hodge microsheaves on cotangent bundles and plumbings
T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash
Pith's one-line read The paper defines Hodge microsheaves and shows they recover the classical Hodge structure on chains of based loop spaces and the Koszul duality of Ginzburg algebras for A_n plumbings.
desk verdict New Hodge microsheaf framework with real applications; the A_n plumbing proof has a checkable gap about whether the needed unipotent local systems actually belong to the chosen saturated category. 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 engine is a saturated mixed structure on microsheaf categories. For a Lagrangian core of Fourier type, the paper assigns to each stratum a saturated non-full subcategory of mixed Hodge modules, then glues these with Fourier transforms and specialization functors to form $\mu\mathrm{MC}_L(X)$; saturation means $\bigoplus_d \mathrm{Hom}_{\widehat{D}}(M, N(d)) \cong \mathrm{Hom}_D(F(M), F(N))$, so mixed Hodge gradings survive as Adams gradings on endomorphism algebras. Hodge wrapping is then computed by taking an inductive system of mixed Hodge modules whose underlying sheaves are the images of the wrapped object under successive Reeb-flow quantizations.
What would settle it
Compute the Hodge endomorphism algebra of the wrapped cotangent fiber for $\mathbb{P}^2$ under two different saturated systems: if the resulting Hodge numbers or weights differ, the claimed independence of the choice fails. Alternatively, test the predicted degree-weight pairs $(0,0),(1,2),(4,6),(5,8),\dots$ for the homology of the based loop space of $\mathbb{P}^2$ against an independent bar-construction calculation.
Extended reading notes
Core claim
The central claim is that a saturated category of Hodge microsheaves $\mu\mathrm{MC}_L(X)$ can be constructed for any holomorphic exact symplectic manifold with a Lagrangian core $L$ of Fourier type, by gluing mixed Hodge modules along normal-bundle strata with Fourier and specialization functors. Saturation guarantees that the endomorphism algebra of a Hodge-lifted object carries an Adams-type grading equal to the total Tate-twisted Hom in the mixed category. The paper then proves two applications: for $X=\mathbb{P}^n$, the saturated Hodge endomorphism algebra of the Hodge-wrapped cotangent fiber matches the classical mixed Hodge structure on the chains of the based loop space; for $A_n$ plumbings of $T^*\mathbb{P}^1$, the Adams-graded Ginzburg algebra is isomorphic to the Hodge endomorphism algebra of the lifted cocores, giving a mixed-geometric proof of the known Koszul duality.
Load-bearing premise
The load-bearing premise is that a chosen saturated system of mixed Hodge modules yields canonical Hodge structures on the wrapped objects, with the projective-space loop-space theorem additionally relying on an external preprint that supplies the explicit geodesic-flow times and constructibility of the flow images.
Editorial extensions
If this is right
- For $X=\mathbb{P}^n$, the wrapped cotangent fiber admits a Hodge lift whose saturated endomorphism algebra reproduces the classical mixed Hodge structure on chains of the based loop space, giving a microlocal construction of that structure.
- For $A_n$ plumbings of $T^*\mathbb{P}^1$, the lifted cocores give an Adams-graded endomorphism algebra isomorphic to the Ginzburg algebra, so the known Koszul duality acquires a mixed-geometric explanation.
- Whenever an object in the saturated system admits an inductive system of mixed Hodge modules along Reeb-flow images, its Hodge wrapping exists and is computable as a colimit.
- Saturated mixed structures can be defined on microsheaf categories for Lagrangian cores of Fourier type, which implies that Adams gradings in such geometric settings come from mixed Hodge theory.
- For the relative core of $A_n$ plumbings, the category $\mathcal{O}$ is recovered from Hodge microsheaves and is Koszul, matching the symplectic-duality picture.
Reading between the lines
- If the saturated-system choice can be shown canonical, the Hodge microsheaf category would become an intrinsic invariant of the holomorphic exact symplectic manifold, not just of the chosen core and auxiliary subcategories; a natural first test is comparing $\mathbb{P}^n$ weights across different saturated systems.
- The same gluing prescription may extend to other conical symplectic resolutions, where it would give mixed-geometric proofs of Koszul duality for category $\mathcal{O}$ beyond the $A_n$ plumbing case; the paper's relative-core example already points in that direction.
- The dependence on an explicit geodesic-flow sequence suggests a general principle: constructibility of the flow kernels controls the existence of Hodge wrappings, so testing the paper's complex-wrapping conjecture on other Fano manifolds would delimit where the method applies.
- The authors' motivation suggests a Hodge-brane version of Fukaya categories in which local systems are replaced by variations of Hodge structure; the sheaf-theoretic route taken here may be the only currently feasible way to make that idea precise.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper introduces a Hodge-theoretic enhancement of the category of microsheaves for holomorphic exact symplectic manifolds whose Lagrangian cores are of Fourier type. The construction proceeds by gluing categories of monodromic mixed Hodge modules via Fourier transforms, following the microlocal Riemann–Hilbert correspondence of Côté–Kuo–Nadler–Shende. The paper also defines a notion of Hodge wrapping and applies it to two concrete settings. First, for X = P^n, it claims that the saturated Hodge endomorphism algebra of the Hodge-wrapped cotangent fiber recovers Hain's mixed Hodge structure on the chains of the based loop space (Theorem 6.6). Second, for A_n plumbings of T*P^1, it constructs Hodge lifts of microlocal skyscraper sheaves and proves (Corollary 8.11) that the Adams-graded Hodge endomorphism algebra of the lifted cocores is quasi-isomorphic to the Ginzburg dga, giving a mixed-geometric proof of Etgü–Lekili's Koszul duality. The proofs rely on an explicit and lengthy appendix that describes the relevant microlocal skyscrapers and their Hodge lifts.
Significance. If the main theorems are correct, the paper would be a significant step toward a microlocal Hodge theory for Weinstein manifolds, providing Hodge structures on wrapped Fukaya categories and giving a conceptual explanation of Adams gradings in symplectic Koszul duality. The explicit construction of microlocal skyscraper sheaves and their Hodge lifts for A_n plumbings is a valuable technical contribution. The paper also formulates a number of conjectures that will likely guide future work. However, the foundational construction of Hodge microsheaves is only sketched, and the main applications depend on the choice of an auxiliary saturated system whose properties are not fully verified; these issues currently weaken the claims.
major comments (4)
- [§4.4, Example 4.9, §9.6] The category µM^c_L(X) depends on the choice of a saturated system (Definition 4.7), and the paper does not prove independence of this choice. For the A_n plumbing, the chosen saturated model is the category M of Example 2.19(2), which is the pretriangulated closure of a category generated by C_{P^1}[1] and skyscraper sheaves C_i, with morphism spaces spanned by id, [P^1], u_i, ι_i. The appendix (Lemma 9.74) asserts that the unipotent monodromic Hodge modules A_s, B_s, P_s, Q_s and the morphisms of Definition-Lemma 9.65 admit Hodge enhancements inside M(V), the restriction of M to an affine chart. This is not verified, and in fact appears false as stated: on the affine chart V, the fundamental class [P^1] restricts to zero, and the generator set has no morphism whose cone could produce the nontrivial local system A_2[1], which is an extension of C_W by C_W with extension class in H^1(W). Since the Hodge wrapping H^{∞,H}_j is built from these objects and the saturatedness of M is used to identify its Hodge endomorphism algebra with the Ginzburg algebra (Corollary 8.11), this is a load-bearing gap.
- [§9.6, Lemmas 9.74–9.78] The verification that the required Hodge structures exist is carried out by direct diagrams and ‘one can check’ statements rather than by a systematic proof that the gluing construction restricts to the saturated subcategories M(V). In particular, Lemma 9.76 claims that the morphisms f_s of Definition-Lemma 9.65 lift to Hodge morphisms A_s → A_s(−1), etc., but it is not shown that these lifts lie in the Hom-spaces of M(V) as defined in Example 2.19(2). If any of these lifts or morphisms is not in M, then the object H^{∞,H}_j constructed in Theorem 9.82 does not lie in µM^C(X_Γ), and the Adams grading of the algebra B computed in Lemma 8.6 and Corollary 9.83 need not coincide with the Ginzburg algebra. A complete proof needs either to show directly that these objects and morphisms are in the saturated model, or to construct a different saturated system that contains them while preserving the Adams-grading computations.
- [§6.3, Theorem 6.6] The proof of Theorem 6.6 depends on the unpublished preprint [Ara] for the explicit sequence of geodesic-flow times and the constructibility of Φ_{t_i}(C_x) on P^n. The paper says that the argument of [Ara] can be adapted, but it does not provide the adaptation. As a result, Theorem 6.6 is conditional on an external result that has not yet appeared in a peer-reviewed venue. The manuscript should either include a complete proof of the needed statement or clearly state the theorem as conditional on [Ara].
- [§4.4] The gluing construction of the category µMHM_L(X) is only sketched. While the equivalence with µCsh_L(X) in the sheaf case is justified by comparison with the gluing description of [CKNSa], the Hodge case is described by the sentence ‘By replacing Sh with √MHM, we define µMHM_L(X)’. It is not proved that the Fourier transforms and specialization functors on monodromic mixed Hodge modules (Section 3.3) are compatible with the diagram defining the gluing, nor that the resulting category is independent of the choices of Darboux coverings and identifications. Since this construction underlies all of the paper's results, a full and rigorous definition is needed before the main theorems can be considered established.
minor comments (5)
- [§2.1] The notation (d) for the Tate twist in Definition 2.1 is not used consistently; throughout the paper the twist is denoted (1) or (1/2). Please unify the notation.
- [§3.3] In equation (3.16), the statement ‘FL^2 = id’ should be clarified: the square is the identity as an autoequivalence of the square-root category, not as a functor on the original category.
- [§5.3, Lemma 5.10] In the displayed equation of the proof, the interchange of colimit and direct sum is not justified; a brief justification or a reference would help.
- [§9.6, Theorem 9.82] The expression (F^0F^{-1})|_{m_j} ∩ (W_1F^{-1})|_{m_j} is terse; it would be helpful to spell out how the Hodge and weight filtrations on the nearby cycle ψ_{m_j}F^{-1} are induced from the filtrations on F^{-1}.
- [§9 (Appendix)] The appendix is very long (over sixty pages) and is organized into many subsubsections; a short introductory overview of the strategy and a dependency diagram among the lemmas would greatly improve readability.
Circularity Check
No significant circularity: the main Hodge computations are independent comparisons against Hain and Etgü–Lekili; the noted issues are reliance on [Ara] and unproven closure of the saturated model, not circular reductions.
full rationale
The paper's construction is definitionally honest. Definition 4.7 introduces a 'saturated system' as a chosen non-full subcategory, and Lemma 4.8 derives saturation of the glued category. The equality (8.2), Hom_{µMC}(H∞,H, ⊕_s H∞,H(s/2)) ≃ End_{µshC}(H∞), is exactly the saturation property of Definition 2.16, so it holds by definition; it is not the paper's prediction, but the device that lets the Tate-twist grading be compared with ordinary endomorphisms. The actual theorems are not forced by this definition. Theorem 6.6 computes the wrapped endomorphism grading from the Hodge weights of the hyperplane and fundamental classes in Arai's twisted complex, then independently computes Hain's bar grading for P^n; the two lists agree. The Tate twists in (6.4) are stated as consequences of those weights, not fitted to (6.7). Similarly, the A_n Koszul duality chapter computes the Adams-graded algebra B from the Hodge lifts Bℓ_j and H∞,H_j, whose half-Tate twists are determined by the Fourier-transform gluing in Lemmas 9.74–9.78, and then identifies B with GΓ via the He–Wu formalism and Etgü–Lekili's known computation. The identification with GΓ uses [EL17] as an external benchmark, so the argument is not circular. The main risks are completeness rather than circularity: Theorems 6.5 and 6.6 depend on [Ara] (an unpublished preprint); Lemma 9.74 asserts that the models A_s, B_s, P_s, Q_s lie in the specific saturated model M(V) without giving a complete proof; and the paper does not establish independence of the chosen saturated system. These gaps could undermine the theorems, but they do not exhibit a reduction of a claimed prediction to its inputs. Self-citations [Sai22, Sai24] supply the Hodge Fourier transform; they are prior tool-building results with stated assumptions and are not equivalent to this paper's outputs.
Assumptions & free parameters
free parameters (4)
- Saturated model M^c_2 for P^n =
objects C_{P^n}[n], C_H[n-1], C_x and specified morphism subspaces (Example 2.19.3)
- Saturated model for A_n plumbing (Example 4.9) =
the saturated model M from Example 2.19.2 for each P^1 with three distinct points l,m,r
- Geodesic-flow times and metric for P^n (from Arai) =
a sequence 0=t0<t1<t2<... and the Fubini–Study metric g
- Microstalk basepoints m_j =
points m_j on the j-th P^1 in the core
assumptions (8)
- standard math Saito's theory of mixed Hodge modules, including six operations and the decomposition theorem (Theorem 2.12)
- standard math Fourier transform of monodromic mixed Hodge modules (Theorem 3.7) is an equivalence lifting the D-module Fourier transform
- domain assumption Ganatra–Pardon–Shende equivalence between partially wrapped Fukaya categories and microsheaves (Theorem 4.3)
- ad hoc to paper The gluing of monodromic Hodge modules via Fourier transforms yields the category of Hodge microsheaves equivalent to the canonical microsheaf category (Section 4.4)
- domain assumption Arai's theorem: for P^n with Fubini–Study metric, geodesic flow images of C_x are constructible at discrete times with explicit twisted complex presentation
- standard math He–Wu Adams Koszul duality theorem (Theorem 7.4)
- domain assumption Etgü–Lekili [EL17] identification of End_F(⊕P^1_v) ≃ A_Γ and End_W(⊕T*_v) ≃ G_Γ for ADE plumbings
- domain assumption Identification of µsh_{C'}(X_Γ) with a block of category O following [BLPW12, Example 4.11]
invented entities (3)
-
Hodge microsheaf category µMHM_L(X)
independent evidence
-
Hodge wrapping
independent evidence
-
Saturated system
Cite this review
Pith. "Pith review of Hodge microsheaves on cotangent bundles and plumbings." pith.science (2026). https://pith.science/paper/WBT63V2O
@misc{pith2026250204148,
author = {Pith},
title = {Pith review of: Hodge microsheaves on cotangent bundles and plumbings},
year = {2026},
howpublished = {\url{https://pith.science/paper/WBT63V2O}},
note = {Machine review of arXiv:2502.04148}
}
read the original abstract
We introduce and study the category of Hodge microsheaves which is a Hodge-version of the category of microsheaves for a certain class of holomorphic exact symplectic manifolds. We then study Hodge-theoretic version of wrapped sheaves and discuss applications in topology and representation theory. Namely, we study (1) Hain's Hodge structures on the cohomology of based loop spaces of algebraic varieties, and (2) the Koszul duality of Ginzburg algebras by Etg\"u-Lekili from a mixed geometric perspective.
Reference graph
Works this paper leans on
-
[7]
Deformation quanti zation and per- verse sheaves
[GS] Sam Gunningham and Pavel Safronov. Deformation quanti zation and per- verse sheaves. arXiv:2312.07595. [Gui16] St´ ephane Guillermou. Quantization of exact Lagra ngian submanifolds in a cotangent bundle, lectures at the 2016 summer school “symp lectic topol- ogy, sheves, and mirror symmetry. available at the author’s webpage ,
arXiv 2016
-
[10]
Wrapped microlocal sheaves on pairs of p ants
[Nad] David Nadler. Wrapped microlocal sheaves on pairs of p ants. arXiv:1604.00114. [NS] David Nadler and Vivek Shende. Sheaf quantization in We instein sym- plectic manifolds. arXiv:2007.10154. [Rid13] Laura Rider. Formality for the nilpotent cone and a d erived Springer correspondence. Adv. Math. , 235:208–236,
arXiv 2007
-
[12]
On the Nori and Hodge realisations of Voe vodsky ´ etale motives
[Tub] Swann Tubach. On the Nori and Hodge realisations of Voe vodsky ´ etale motives. arXiv:2309.11999. Tatsuki Kuwagaki: Department of Mathematics, Kyoto Univer sity, Kitashirakawa Oiwake- cho, Sakyo-ku, Kyoto 606-8502, Japan. E-mail address : tatsuki.kuwagaki.a.gmail.com Takahiro Saito: Department of Mathematics, Chuo Universit y, 1-13-27 Kasuga, Bunkyo...
-
[251]
On $V$-filtration, Hodge filtration and Fourier transform
1986 . G´ eom´ etrie et analyse microlocales. [CD] Qianyu Chen and Bradley Dirks. On V-filtration, Hodge fil tration and Fourier transform. arXiv:2111.04622. [CDRGG24] Baptiste Chantraine, Georgios Dimitroglou Rize ll, Paolo Ghiggini, and Roman Golovko. Geometric generation of the wrapped Fukaya c ategory of Weinstein manifolds and sectors. Ann. Sci. ´Ec. N...
work page Pith review arXiv 1986
-
[1973]
The microlocal Riemann-Hilbert correspondence for complex contact manifolds
[CKNSa] Laurent Cˆ ot´ e, Christopher Kuo, David Nadler, and Vivek Shende. The mi- crolocal Riemann–Hilbert correspondence for complex cont act manifolds. arXiv:2406.16222. 109 [CKNSb] Laurent Cˆ ot´ e, Christopher Kuo, David Nadler, and Vivek Shende. Per- verse microsheaves. arXiv:2209.12998. [EL17] Tolga Etg¨ u and YankıLekili. Koszul duality pattern s ...
-
[1994]
With a chapter in French by Christian Houzel, Corrected reprint o f the 1990 original. [Kuo23] Christopher Kuo. Wrapped sheaves. Adv. Math. , 415:Paper No. 108882, 71,
work page 1990
-
[2012]
[GMW] Benjamin Gammage, Michael McBreen, and Ben Webster. H omologi- cal mirror symmetry for hypertoric varieties ii (with an app endix written jointly with laurent cˆ ot´ e and justin hilburn). arXiv:1903.07928. [GPS24a] Sheel Ganatra, John Pardon, and Vivek Shende. Micr olocal Morse theory of wrapped Fukaya categories. Ann. of Math. (2) , 199(3):943–1042,
-
[2014]
Guillermou-Kashiwara-Schapira kernels of geodesic flows
108 [Ara] Takumi Arai. Guillermou–Kashiwara–Schapira kernel s of geodesic flows. arXiv:2502.19978. [BBD+15] C. Brav, V. Bussi, D. Dupont, D. Joyce, and B. Szendr˝ oi. S ymmetries and stabilization for sheaves of vanishing cycles. J. Singul. , 11:85–151,
Show all 12 references
-
[2015]
Wrapped Fukaya categ ory of plumb- ings
[KL] Dogancan Karabas and Sangjin Lee. Wrapped Fukaya categ ory of plumb- ings. arXiv:2405.10783. [KR08] Masaki Kashiwara and Rapha¨ el Rouquier. Microlocal ization of rational Cherednik algebras. Duke Math. J. , 144(3):525–573,
-
[2019]
Duality between Lagr angian and Legen- drian invariants
[EL23] Tobias Ekholm and YankıLekili. Duality between Lagr angian and Legen- drian invariants. Geom. Topol., 27(6):2049–2179,
-
[2021]
MHM project
111 [SS] Claude Sabbah and Christian Schnell. MHM project. available on author’s webpage. [SV19] Jake P. Solomon and Misha Verbitsky. Locality in the F ukaya category of a hyperk¨ ahler manifold.Compos. Math., 155(10):1924–1958,
1924
-
[2024]
Hypertoric Fukaya categories and categories O
[CGH] Laurent Cˆ ot´ e, Benjamin Gammage, and Justin Hilburn . Hypertoric Fukaya categories and categories O. arXiv:2406.01379. [Che73] Kuo-tsai Chen. Iterated integrals of differential fo rms and loop space homology. Ann. of Math. (2) , 97:217–246,
Reviewed August 8, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.