{"id":"fbc5536f-d750-48ec-b103-1ce81016da55","arxiv_id":"2504.15868","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Regular semisimple Lusztig varieties are geometrically well-behaved and flatly degenerate to Hessenberg varieties, proving the Abreu-Nigro conjecture and the affineness of Deligne-Lusztig cells.","lead":"Regular semisimple Lusztig varieties, subvarieties of flag manifolds attached to Weyl group elements, are shown to be normal, Cohen-Macaulay, and to have affine open cells and vanishing cohomology. The paper also constructs a flat degeneration of these varieties into Hessenberg varieties, proving a 2024 conjecture of Abreu and Nigro and settling a 1976 question on Deligne-Lusztig cells.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Flatness of the degeneration family bY_w -> bG (Theorem 5.13) is asserted from fiber dimension alone; Miracle Flatness needs Cohen-Macaulay, which is unproved, so the Ehresmann step in Corollary 5.16 is not secured.","rationale":"The paper's first part (singularities, vanishing, Frobenius splitting) is well-supported by detailed proofs and appears sound. The degeneration theorem is the central new bridge, and the reader correctly isolates its flatness justification as the weakest point. My reading of Section 5.3 confirms that the asserted criterion is false as stated: Miracle Flatness needs Cohen-Macaulay, and the construction does not yield it for free. The fiber computation in Lemma 5.14 is also only sketched, but even if correct, it only gives the set-theoretic fiber; flatness is what upgrades it to a genuine degeneration. Since this is an omitted hypothesis in the central claim, the appropriate verdict remains CONDITIONAL: the main conclusions are plausible and likely repairable, but the paper must supply a proof of flatness (or a different argument for the diffeomorphism). I therefore keep the reader's verdict unchanged.","tokens_in":42,"tokens_out":5519,"duration_ms":477791,"concrete_test":"Check flatness of (5.12) directly in local coordinates. For G = SL_2 and w = s (or a rank-2 example), write explicit equations for bY_w from diagram (5.11) in an affine chart of bG x B over the exceptional divisor bE; then verify that O_{bY_w} is a Cohen-Macaulay O_{bG}-module by computing depth over the local ring of the base, or compute Tor^1 over a point in bE. A vanishing Tor computation in Macaulay2 would settle whether the family is flat; if Tor is nonzero, Theorem 5.13's flatness claim fails. Alternatively, prove CM of bY_w by exhibiting a regular sequence generating the ideal of bY_w in the smooth ambient Bl_{bE x Delta(B)}(bG x B x B) locally at points of the exceptional fiber.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 5.13 concludes: \"Since bG is smooth and all the fibers have the same dimension ..., (5.12) is flat.\" This is not a valid flatness criterion: equidimensionality of fibers plus a smooth base does not imply flatness. The standard Miracle Flatness theorem requires the total space bY_w (or every fiber) to be Cohen-Macaulay, and the paper neither proves bY_w is Cohen-Macaulay nor cites a variant avoiding this. The total space is obtained as a fiber product of blowups in diagram (5.11), so Cohen-Macaulayness is not automatic, and the special fiber over bE is only described set-theoretically in Lemma 5.14. For smooth w, the assertion that the family is smooth also presupposes flatness; fiberwise smoothness plus a smooth base does not imply flatness either. If flatness fails, the degeneration theorem 5.13 is not established, and the Ehresmann theorem used in Corollary 5.16 (and hence the proof of the Abreu-Nigro conjecture) has no basis. This is the load-bearing step connecting the GKM-graph comparison in Theorem 5.12 to the diffeomorphism result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies regular semisimple Lusztig varieties Y_w(s) in a flag variety. The first half proves structural results: these varieties are normal, Cohen-Macaulay, of pure dimension l(w), have rational singularities, and enjoy vanishing theorems for ample and nef line bundles in arbitrary characteristic, via anti-canonical divisors, Kawamata-Viehweg vanishing, and Frobenius splitting. It also proves that Lusztig cells and Deligne-Lusztig cells are affine. The second half relates Lusztig varieties to Hessenberg varieties. For smooth w, the paper defines a Hessenberg space H_w from the tangent space of X_w at the identity, proves that Y_w(s) and X_{H_w}(s) have the same GKM graphs, and constructs a family bY_w over a blowup bG of G at the identity that is claimed to be a flat degeneration of the Lusztig family over G^rs to the Hessenberg family over P(g)^rs. From this degeneration, the paper derives a diffeomorphism result (Corollary 5.16) proving and generalizing the Abreu-Nigro conjecture.","tokens_in":33240,"tokens_out":24770,"duration_ms":236726,"significance":"If the main degeneration theorem is correct, the paper resolves a conjecture of Abreu and Nigro and extends it to all Lie types; it also provides a uniform geometric explanation for the known cohomological relationship between Lusztig and Hessenberg varieties. The first part of the paper is strong and largely self-contained: the singularity results in Section 2, the anti-canonical computation in Proposition 3.1, the vanishing arguments in Sections 3 and 4, and the affineness results for Lusztig and Deligne-Lusztig cells are substantial and, as far as I can see, correctly argued. The GKM comparison in Theorem 5.12 is a natural extension of the type-A results of Abreu-Nigro and is convincingly sketched. The degeneration construction itself is original and potentially important. However, the flatness of the degeneration family is asserted rather than proved, and the special-fiber computation is only sketched; since the diffeomorphism conclusion depends directly on these points, the central claim is not yet fully established.","major_comments":[{"comment":"The flatness of bY_w -> bG is not established. The final paragraph of Section 5.3 says: 'Since bG is smooth and all the fibers have the same dimension ..., (5.12) is flat.' This is not a valid flatness criterion. The standard Miracle Flatness theorem requires the total space bY_w (or, in some variants, the fibers) to be Cohen-Macaulay; bY_w is constructed as a fiber product of blowups in diagram (5.11), so its Cohen-Macaulayness is not automatic and is not proved. The parenthetical claim that the family is smooth when w is smooth also requires flatness (or a proof that bY_w is smooth and then an application of Miracle Flatness); smooth fibers alone do not imply that a morphism is smooth. Since Corollary 5.16 applies Ehresmann's theorem to this family, and since Conjecture 1.7 is derived from that corollary, this gap is load-bearing. Please prove flatness directly, for example by proving that bY_w is Cohen-Macaulay or smooth, or by showing that the fibers have constant Hilbert polynomial over the connected base bG.","section":"Section 5.3, proof of Theorem 5.13, equation (5.12)"},{"comment":"The identification of the special fiber over bE is only sketched. The proof asserts several nontrivial identifications: that the map from bE×B to P_{bE×B}(pr_1^*O_{bE}(-1) ⊕ pr_2^*T_B) factors through pr_2^*T_B, that the relevant preimage of N_{\\Delta(B)/B×B} is C_{\\Delta(B)/X_w} ≅ G×^B H_w/b, and that the fiber is therefore the fiber product (5.2) with H=H_w. These claims are not derived; the phrase 'one can easily check, for example using one-parameter families' is not a substitute for a normal-cone computation. Because this lemma identifies the closed fiber of the degeneration as the Hessenberg variety X_{H_w}(s), it is essential for the Ehresmann comparison in Corollary 5.16. I recommend that the proof of Lemma 5.14 be expanded into a complete computation.","section":"Section 5.3, Lemma 5.14"}],"minor_comments":[{"comment":"There are typographical errors: 'Deline-Lusztig' should be 'Deligne-Lusztig' in the abstract and in the heading of Corollary 3.12.","section":"Abstract and Section 3.3, Corollary 3.12"},{"comment":"The word 'Gorentein' should be 'Gorenstein'.","section":"Section 2.3, proof of Theorem 2.12"},{"comment":"The notation Y_e(s) = {wB : w∈W} uses the symbol w both for the fixed Weyl element and for the running variable; this is confusing and should be rephrased, for example as {vB : v∈W}.","section":"Section 2.3, Remark 2.8(1)"},{"comment":"The proof of Theorem 5.12 compares T-invariant curves, but the conclusion is an isomorphism of graded W-representations. Please state explicitly that the GKM graphs are compared in the labeled sense, with vertices labeled by elements of W and with the natural W-action on the graph; the underlying unlabeled graph alone would not determine the dot action.","section":"Section 5.2, Theorem 5.12"}],"recommendation":"major_revision","confidential_remarks":"The paper contains substantial correct material, and the main degeneration construction is promising. The only serious obstacle I see is the unproved flatness/smoothness of bY_w -> bG in Theorem 5.13 and the terse proof of Lemma 5.14. Both are localized and likely fixable within the scope of the manuscript, so I recommend major revision rather than rejection. If the authors can prove that bY_w is Cohen-Macaulay (or smooth for smooth w) and supply the missing normal-cone details, I would expect the paper to be publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read Brosnan–Hong–Lee. The good news first: Sections 2–4 are a solid contribution. They prove regular semisimple Lusztig varieties are normal, Cohen–Macaulay, rational singularities, and pure dimension, with arbitrary-characteristic vanishing for ample line bundles (nef in char 0 or large p) via Bott–Samelson resolutions, an anti-canonical boundary divisor, Kawamata–Viehweg, and Frobenius splitting. The affineness of Lusztig cells, and especially of Deligne–Lusztig cells, answers an honest open question from 1976. The GKM graph comparison in Theorem 5.12 is clean, and the map from smooth elements of W to Hessenberg spaces via tangent cones is a natural generalization of Abreu–Nigro. Citations look appropriate, with prior type-A work credited properly.\n\nThe soft spot is exactly where the stress-test note lands. In Theorem 5.13, flatness of bY_w -> bG is concluded from 'bG smooth and all fibers same dimension.' That is not a valid criterion. Miracle Flatness requires the total space or the fibers to be Cohen–Macaulay, and neither is proved. bY_w is a fiber product of blowups in diagram (5.11), so CM is not automatic. Lemma 5.14 only describes the special fiber set-theoretically, not its scheme structure. Without flatness, the Ehresmann step in Corollary 5.16—and hence the proof of the Abreu–Nigro conjecture—is not secured. This is load-bearing, not cosmetic.\n\nI would not desk-reject. The first half alone is a refereed paper, and the degeneration construction is plausible and likely fixable by proving bY_w is CM or by building the family differently. The authors should be asked for a rigorous flatness proof and a scheme-theoretic description of the special fiber. The cited sentence reads like a misremembered criterion, not a hopeless gap.\n\nWho gets value: algebraic geometers working on flag varieties, Hessenberg varieties, Frobenius splitting, and the representation-theoretic side. Bring it to reading group; the gap is instructive.","headline":"Strong first half, but Theorem 5.13's flatness claim is not justified—worth peer review with a required fix.","tokens_in":33728,"tokens_out":3645,"would_cite":true,"duration_ms":37307,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14M15","14F17","14B05","14L30","14D06","14M17"],"pacs":[],"model":"deepseek-v4-flash","headline":"Regular semisimple Lusztig varieties degenerate flatly to Hessenberg varieties, proving a type A homeomorphism conjecture and generalizing it to all Lie types.","keywords":["Lusztig varieties","Hessenberg varieties","Schubert varieties","Bott-Samelson resolutions","Frobenius splitting","cohomology vanishing","GKM graphs","flat degeneration"],"falsifier":"Compute the exceptional fiber of bY_w → bG in a small singular case, such as G = GL_4 with w = 4231 where the tangent cone is a quadric cone, and compare it with the Hessenberg scheme X_{H_w}(s); if they differ, Theorem 5.13(2) fails. Independently, test whether the total space bY_w is Cohen-Macaulay: if not, the flatness criterion invoked at the end of Section 5.3 does not apply, and the diffeomorphism conclusion has no proof.","tokens_in":32780,"feed_emoji":"📐","tokens_out":9603,"duration_ms":82690,"temperature":0.7,"pith_summary":"This paper studies Lusztig varieties, subvarieties of the flag manifold defined by the condition that $g^{{-1}}$sg lies in a Bruhat double coset, for s regular semisimple. It claims these varieties are normal, Cohen-Macaulay, of pure dimension ℓ(w), with rational singularities, and that ample line bundles on them have no higher cohomology in any characteristic. The main structural result is stronger: there is a smooth projective family over a blowup of G whose fibers over regular semisimple elements are Lusztig varieties and whose fibers over the exceptional divisor are Hessenberg varieties. That flat degeneration makes Hessenberg varieties the linearized limits of Lusztig varieties, and it implies that two smooth elements of the Weyl group with the same Schubert tangent space at the identity give diffeomorphic Lusztig varieties. In type A this proves the conjectured homeomorphism, and the statement holds in arbitrary Lie type.","feed_headline":"Flat family joins Lusztig and Hessenberg varieties","feed_subtitle":"The degeneration proves a type A homeomorphism conjecture and extends it to all Lie types.","key_machinery":"The central mechanism is a deformation-to-normal-cone family built from the universal Lusztig family. One starts with the fiber product Y_w ⊂ G×B defined by $g^{{-1}}$xg ∈ BwB, pulls it back along the blowup bG of G at the identity, and blows up the diagonal inside B×B along the exceptional locus; the exceptional fiber becomes the projectivized normal cone, which is exactly the Hessenberg scheme X_{H_w}(s). The Hessenberg space H_w is $π_b^{{-1}}$(C_w), where C_w is the tangent cone of the Schubert variety X_w at the identity. On the cohomology side, the GKM graphs of Y_w(s) and X_{H_w}(s) coincide for smooth w, transferring the Weyl-group dot action from one variety to the other. The singularity results are carried by Bott-Samelson resolutions, whose boundary is a simple normal crossing anti-canonical divisor; in positive characteristic that divisor yields Frobenius splittings, and in characteristic zero it feeds a Kawamata-Viehweg vanishing argument.","core_discovery":"The paper's central discovery is that regular semisimple Lusztig varieties and regular semisimple Hessenberg varieties are not merely cohomologically parallel: they are the two ends of a single flat family. After blowing up G at the identity and deleting the proper transform of the non-regular locus, the paper constructs a smooth projective morphism bY_w → bG whose fiber over a regular semisimple element s is the Lusztig variety Y_w(s), and whose fiber over a point [s] in the exceptional divisor is the Hessenberg variety X_{H_w}(s), with H_w built from the tangent cone of the Schubert variety X_w at the identity. For smooth w this family is smooth, so a standard topological argument makes all fibers diffeomorphic; two smooth Weyl group elements with the same Schubert tangent space therefore yield diffeomorphic Lusztig varieties. That proves the type A homeomorphism conjecture and extends it to all Lie types. Along the way the paper establishes that regular semisimple Lusztig varieties are normal, Cohen-Macaulay, of pure dimension ℓ(w), with rational singularities, that their ample line bundles have vanishing higher cohomology in all characteristics, and that Lusztig cells and Deligne-Lusztig cells are affine.","pith_inferences":["One consequence the paper leaves implicit: the degeneration family is defined for all w, not only smooth ones, so the same construction may transfer intersection-cohomology or Hodge-theoretic invariants from Hessenberg schemes to singular Lusztig varieties.","The tangent-cone map w ↦ H_w is purely combinatorial; classifying which Hessenberg spaces arise from smooth Schubert varieties (the paper notes type C_3 exceptions) would give a concrete test of how sharp the degeneration theorem is.","If the flatness gap is repaired, the family should put the monodromy actions on both sides into a single local system over bG, making the dot-action comparison a statement about monodromy rather than about GKM graphs alone.","The affineness of Deligne-Lusztig cells may lead to new cohomological vanishing for the representations they carry; testing whether the same argument works for non-reduced words or other Frobenius twists is a natural next step."],"forward_implications":["Smooth Weyl group elements w and w′ with the same Schubert tangent space at the identity yield diffeomorphic regular semisimple Lusztig varieties for any regular semisimple elements; in type A this is the conjectured homeomorphism.","Ample line bundles on every regular semisimple Lusztig variety have vanishing higher cohomology in arbitrary characteristic, and nef line bundles do too in characteristic zero or sufficiently large characteristic.","Lusztig cells are affine, as are Deligne-Lusztig cells, resolving an open question from the first papers on Deligne-Lusztig varieties.","For smooth w and dominant λ, the Weyl group acts on the weight spaces of H^0(Y_w(s), L_λ) with equal dimensions along each orbit.","Cohomology vanishing and restriction surjectivity transfer to all regular semisimple Hessenberg varieties in type A, extending earlier results for weak Fano cases."],"supporting_citations":[{"why":"Introduces Lusztig varieties as fibers of a family used to construct character sheaves.","marker":"[40]"},{"why":"Establishes the type A comparison and conjecture on homeomorphism types that the paper proves and generalizes.","marker":"[3]"},{"why":"Introduces Hessenberg varieties and proves their smoothness for regular semisimple elements.","marker":"[19]"},{"why":"Provides the Schubert variety results and Bott-Samelson resolution arguments that are extended to Lusztig varieties.","marker":"[11]"},{"why":"Supplies the deformation-to-normal-cone construction used to build the degeneration family.","marker":"[24]"},{"why":"Supplies the Frobenius-splitting criteria used to prove cohomology vanishing in positive characteristic.","marker":"[44]"},{"why":"Introduces Deligne-Lusztig varieties and raises the affineness question settled here.","marker":"[20]"},{"why":"Describes line bundles on relative Bott-Samelson varieties, used to prove ampleness of the boundary divisors.","marker":"[6]"},{"why":"Describes line bundles on Bott-Samelson varieties and their ampleness, used to show boundary divisors are ample.","marker":"[37]"},{"why":"Introduces the dot action of the Weyl group on Hessenberg cohomology, the representation-theoretic target.","marker":"[56]"}],"fun_headline_variants":["Flat family unifies Lusztig and Hessenberg varieties","Lusztig varieties degenerate to Hessenberg varieties","Conjecture proven: Lusztig and Hessenberg are one family","All Lie types: Lusztig and Hessenberg cohomology agree","Lusztig and Hessenberg: one flat degeneration"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that bY_w → bG is flat: the paper invokes a standard flatness criterion at the end of Section 5.3 that requires a Cohen-Macaulay hypothesis on the total space or fibers, and that hypothesis is not established for bY_w; if flatness fails, the degeneration and the diffeomorphism corollary do not follow.","fun_headline_variants_meta":{"raw":{"variants":["Flat family unifies Lusztig and Hessenberg varieties","Lusztig varieties degenerate to Hessenberg varieties","Conjecture proven: Lusztig and Hessenberg are one family","All Lie types: Lusztig and Hessenberg cohomology agree","Lusztig and Hessenberg: one flat degeneration"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000518,"raw_usage":{"total_tokens":2618,"prompt_tokens":1158,"completion_tokens":1460,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":774,"completion_tokens_details":{"reasoning_tokens":1371}},"tokens_in":774,"tokens_out":1460,"duration_ms":11190,"temperature":1.0,"reasoning_tokens":1371,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:17:11.569553+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the exceptional fiber of bY_w → bG in a small singular case, such as G = GL_4 with w = 4231 where the tangent cone is a quadric cone, and compare it with the Hessenberg scheme X_{H_w}(s); if they differ, Theorem 5.13(2) fails. Independently, test whether the total space bY_w is Cohen-Macaulay: if not, the flatness criterion invoked at the end of Section 5.3 does not apply, and the diffeomorphism conclusion has no proof.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces Lusztig varieties as fibers of a family used to construct character sheaves."},{"cited_title":"Abreu and A","cited_arxiv_id":null,"evidence_quote":"Establishes the type A comparison and conjecture on homeomorphism types that the paper proves and generalizes."},{"cited_title":"De Mari, C","cited_arxiv_id":null,"evidence_quote":"Introduces Hessenberg varieties and proves their smoothness for regular semisimple elements."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Schubert variety results and Bott-Samelson resolution arguments that are extended to Lusztig varieties."},{"cited_title":"Fulton.Intersection theory, volume 2 ofErgebnisse der Mathematik und ihrer Grenzgebiete","cited_arxiv_id":null,"evidence_quote":"Supplies the deformation-to-normal-cone construction used to build the degeneration family."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Frobenius-splitting criteria used to prove cohomology vanishing in positive characteristic."},{"cited_title":"Deligne and G","cited_arxiv_id":null,"evidence_quote":"Introduces Deligne-Lusztig varieties and raises the affineness question settled here."},{"cited_title":"Bhaumik and P","cited_arxiv_id":null,"evidence_quote":"Describes line bundles on relative Bott-Samelson varieties, used to prove ampleness of the boundary divisors."},{"cited_title":"Lauritzen and J","cited_arxiv_id":null,"evidence_quote":"Describes line bundles on Bott-Samelson varieties and their ampleness, used to show boundary divisors are ample."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the dot action of the Weyl group on Hessenberg cohomology, the representation-theoretic target."}],"review_version":1}