{"id":"393f66ad-3ad0-4c93-9e87-aacb7671d95f","arxiv_id":"2411.18273","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For any finite type, generalized Schur algebras are realized as convolution algebras in Borel-Moore homology and equivariant K-theory of Steinberg varieties.","lead":"The paper builds generalized Schur algebras, a family of associative algebras attached to any Lie type, from the geometry of flag varieties and nilpotent cones. It connects these geometric constructions to local Langlands reciprocity, ıquantum groups, and Howe duality.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.11's one-sentence induction asserts an R(G)-splitting that is not established; without it, the freeness/basis of K^G(Zγν) and the injectivity step of Theorem 5.21 are unsupported.","rationale":"The reader's weakest assumption identifies Proposition 5.11 as the main gap, and my stress-test agrees. The proof of Proposition 5.11 is a single sentence invoking induction on w, but the induction requires a splitting of the equivariant K-theory filtration over R(G). This splitting is not a consequence of Lemma 5.5 alone, since R(G) is not a PID and the relevant exact sequences need not split. The cellular fibration lemma, which would provide splitting in the regular-orbit case, is explicitly inapplicable for general γ,ν (Example 5.10). The consequence is that Corollary 5.6's rank computation and Theorem 5.21's injectivity proof both rest on an unproved basis statement. This is not a disagreement with the consensus or an internal inconsistency; it is a genuinely missing argument in a technically central place. The proposed concrete test isolates the smallest case where the known splitting mechanism fails and checks the asserted freeness and basis directly. If the test shows the sequence is non-split, the main theorem requires substantial repair; if it splits, the conditional verdict can be upgraded. Therefore the reader's CONDITIONAL verdict is appropriate, and no change to the verdict is needed at this stage.","tokens_in":47322,"tokens_out":3873,"duration_ms":38295,"concrete_test":"Fix G=GL_3(C), Q_f={γ} with Pγ=B{1,s1}B, and take ν=γ, w=s2 in D_{γγ}, the configuration of Example 5.10. Compute K^G(Z^{≼w}_{γγ}) by an independent method, for instance by presenting Z^{≼w}_{γγ} as a union of two conormal bundles over G/P^w_{γγ} and G/P^{/BD}_{γγ}, and using equivariant localization and Koszul resolutions to determine its structure as an R(G)-module. Then check whether the exact sequence 0→K^G(Z^{≺w}_{γγ})→K^G(Z^{≼w}_{γγ})→K^G(T^*_{O_{γ,w,γ}})→0 splits over R(G) and whether K^G(Z^{≼w}_{γγ}) is free of rank #D_γ^2 with the basis asserted in Proposition 5.11. If the sequence does not split, Proposition 5.11 and the injectivity argument in Theorem 5.21 fail in this case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central equivariant K-theoretic result, Theorem 5.21(2), is proved by showing injectivity of ~ψ: K^G(Zf) → ~Sf and matching ranks. The injectivity argument uses Proposition 5.11 to expand a nonzero element M ∈ K^G(Zf) in the asserted basis {χ^w_{γν}} of K^G(Z_{γν}). Proposition 5.11, however, is proved in one sentence: 'The statements can be proved by induction on w.' The induction must show that each exact sequence 0 → K^G(Z^{≺w}_{γν}) → K^G(Z^{≼w}_{γν}) → K^G(T^*_{O_{γ,w,ν}}) → 0 (from Lemma 5.5) splits as an R(G)-module, and that the chosen classes ι_w(χ) form a basis. The paper asserts a 'split R(G)-homomorphism' K^G(T^*_{O_{γ,w,ν}}) → K^G(Z^{≼w}_{γν}) immediately before Proposition 5.11, but no proof is given that this is a genuine section of the quotient map, nor that the extensions are trivial. This is not a minor formality: R(G) is generally not a PID, so freeness of the submodule and quotient does not force freeness of the middle term. The usual mechanism for splitting, the cellular fibration lemma (Lemma 5.3), is explicitly unavailable for general γ,ν, as Example 5.10 shows: for G=GL_n, Pγ=B{1,s1}B and w=s2 with s1w≠ws1, the relevant fiber is a disjoint union of two cells rather than an affine space. Thus the exact cases needed for Corollary 5.6 and Theorem 5.21 are precisely those where the only known splitting argument fails. If the extension is nontrivial, K^G(Z_{γν}) may fail to be free over R(G), the rank computation in Corollary 5.6 would be invalid, and the injectivity argument in Theorem 5.21 would break. The authors' transparency about the indirect proof is commendable, but Proposition 5.11 is load-bearing and underproved.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs generalized Steinberg varieties Z_f attached to finite W-invariant subsets of the weight lattice of a reductive group G, and proves two main structural results: Theorem 3.11, an isomorphism between the top Borel-Moore homology H(Z_f) and the Schur algebra S_f, and Theorem 5.21, an isomorphism between the equivariant K-group K^{Ĝ}(Z_f) and the affine q-Schur algebra ~S_f. The proof strategy for Theorem 5.21 goes through a Fock-space module ~T_f, realized as K^{Ĝ}(~N_f ×_N ~N), and then identifies the convolution algebra of Z_f with the endomorphism algebra of that module. The paper also derives applications: a Schur-algebra analogue of the local geometric Langlands correspondence, an equivariant K-theoretic realization of quasi-split affine ıquantum groups of type AIII (implicitly, via multiplication formulas from the authors' earlier work), and geometric Howe dualities.","tokens_in":47766,"tokens_out":3357,"duration_ms":34972,"significance":"If the main theorems are correct, the paper gives a genuinely uniform geometric construction of Schur algebras and affine q-Schur algebras for all types, reaching beyond the previously known type-A and type-C cases and recovering the affine Hecke algebra realization as the special case of a single regular W-orbit. The applications are substantial: a Langlands-type reciprocity for generalized Schur algebras, a new route to affine ıquantum groups, and Howe dualities at the level of equivariant K-theory. The paper has clear strengths: the Borel-Moore homology part is supported by explicit upper-triangular formulas and dimension counts (Theorem 3.11), the results are checked against the independent Ginzburg realization of affine Hecke algebras (Theorem 5.12), and there are no fitted parameters or ad hoc axioms. The main gap is concentrated in one proof: Proposition 5.11, on which the rank computation and the injectivity argument of Theorem 5.21 rest, is asserted in a single sentence without establishing the required splitting.","major_comments":[{"comment":"The proof of Proposition 5.11 is one sentence ('The statements can be proved by induction on w') and does not establish the claimed R(Ĝ)-basis of K^{Ĝ}(Z^{≼w}_{γν}). The preceding paragraph asserts a 'split R(Ĝ)-homomorphism' K^{Ĝ}(T^*_{O_{γ,w,ν}}) → K^{Ĝ}(T^*_{γ,w,ν}) but gives no section and no verification that the quotient maps in the exact sequences of Lemma 5.5 split. This is load-bearing: R(Ĝ) is generally not a PID, so freeness of the submodule and of the graded pieces does not force freeness of the middle term, and the standard splitting mechanism, Lemma 5.3, is explicitly unavailable for general γ,ν, as Example 5.10 shows (for G=GL_n, P_γ=B⟨1,s_1⟩B, and w=s_2 with s_1w≠ws_1, the relevant fiber is a disjoint union of two cells). Since Corollary 5.6 and the injectivity step of Theorem 5.21(2) both depend on this basis, a complete proof of Proposition 5.11, or a precise citation to a result that supplies it, is necessary.","section":"§5.7, Proposition 5.11"},{"comment":"The injectivity argument for ~ψ uses Proposition 5.11 to expand a nonzero element M ∈ K^{Ĝ}(Z_f) as M ∈ χ^w_{γν} + K^{Ĝ}(Z^{≺w}_{γν}) and then applies Lemma 5.20 to obtain a nonzero restriction. If the basis assertion of Proposition 5.11 is not available, this expansion and the subsequent rank comparison via Corollary 5.6 fail. The final Cartesian-square argument that identifies End_{R(Ĝ)}(~T_f) ∩ K^{Ĝ}(Z_f)_{loc} with K^{Ĝ}(Z_f) also uses the same freeness/basis structure. The theorem is the central equivariant K-theoretic claim of the paper, so the proof needs to be completed at this point rather than deferred to an induction statement.","section":"§5.12, Theorem 5.21(2)"},{"comment":"The statement 'we provide the geometric construction of a series of generalized Schur algebras of any type' is stronger than what the K-theoretic part proves as written. Section 5.4 assumes that G is connected reductive and that each simple factor of its derived subgroup is simply connected or of type SO_{2d+1}. While this may cover all Weyl types by passing to simply connected covers, the abstract and Theorem 7.1 should state the assumption explicitly, especially because the Langlands-dual formulation in Theorem 7.1 inherits it. Please add a sentence in the introduction and in Theorem 7.1 specifying the class of groups for which the equivariant K-theoretic and Langlands-reciprocity results are proved.","section":"§5.4 and abstract/introduction"}],"minor_comments":[{"comment":"The word 'Cartisian' appears in the headings and text of Lemma 3.3 and Lemma 5.14; it should be 'Cartesian'.","section":"§3.3 and §5.9"},{"comment":"The text 'Fork space ~Tf' appears to be a typo for 'Fock space ~Tf'.","section":"§7.3.1"},{"comment":"The citation '[FL3W220]' in §1.5 does not match the reference list entry '[FL3W20]'; please correct the tag and check the year consistency.","section":"§1.5 and References"},{"comment":"The phrase 'replace U(gl_m) by U(sl_m) (resp, U(gl_n))' is unclear; presumably it should read 'replace U(gl_m) by U(sl_m) (resp. U(gl_n) by U(sl_n))'.","section":"§7.4.2, Theorem 7.5 remark"}],"recommendation":"major_revision","confidential_remarks":"The central issue is Proposition 5.11. The authors are clearly aware that the cellular fibration lemma fails for general γ,ν (Example 5.10), and they state the desired splitting without proof. The rest of the K-theoretic argument is coherent and the rank computations are plausible, but the missing induction is load-bearing. I would not recommend rejection, provided a complete proof of the splitting and freeness is supplied or a precise reference is given. The paper would then be a strong contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe headline: this paper is a serious, technically dense contribution that deserves referee time, but it has a load-bearing gap in the proof of the K-theoretic main theorem. If Proposition 5.11 is fixable, the results are very good; as written, the proof is incomplete.\n\nWhat's new: the paper gives a uniform geometric construction of generalized Schur algebras for any finite W-invariant subset Qf, in both Borel-Moore homology (Theorem 3.11) and equivariant K-theory (Theorem 5.21). This generalizes known type-A results (BLM, Ginzburg-Vasserot, Vasserot) and Su-Wang's type-C case, and it's the first any-type statement at this level of generality. The applications—a Schur-level local geometric Langlands reciprocity, an implicit K-theoretic realization of affine ıquantum groups, and geometric Howe duality—are natural and well-motivated. The authors are transparent about the indirect nature of the arguments and the restriction on G in §5.4.\n\nThe classical-limit part seems solid: Theorem 3.11 follows from upper triangularity plus a dimension count, and the proof is reasonably complete. The soft spot is Theorem 5.21, specifically Proposition 5.11. The proof of that proposition is one sentence: \"The statements can be proved by induction on w.\" But the induction cannot work without proving that the exact sequences of Lemma 5.5 are split as R(G)-modules. The paper simply asserts a \"split R(G)-homomorphism\" without constructing a genuine section. This matters: R(G) is not a PID, so freeness of the submodule and quotient does not imply freeness of the middle term. And Example 5.10 shows the usual cellular fibration mechanism fails in precisely the cases needed. So the basis of K^G(Z_{γν}), the rank computation in Corollary 5.6, and the injectivity step in Theorem 5.21 all rest on this unproved splitting.\n\nI should also note that Application II doesn't derive multiplication formulas for generators from the K-theoretic construction; it imports them from the authors' own [FL3W23]. That's fine if the goal is an implicit realization, but it means the ıquantum group application is not independent of prior work.\n\nBottom line: the paper is worth engaging with seriously. The authors are not sloppy; they know where the difficulties are. But a referee should insist on a complete proof of Proposition 5.11 before accepting. If the splitting can be established, this is a strong paper. As it stands, it's a well-motivated conditional.","headline":"A serious, dense paper with a real advance and one genuinely underproved step (Proposition 5.11) that the authors should be asked to fix before this is accepted.","tokens_in":48348,"tokens_out":3766,"would_cite":false,"duration_ms":34662,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20G43","22E57","20G42"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that generalized Steinberg varieties carry convolution algebras isomorphic to ordinary and affine q-Schur algebras in every finite type.","keywords":["Schur algebras","Steinberg varieties","Borel-Moore homology","equivariant K-theory","affine q-Schur algebras","quasi-split affine ı-quantum groups","local geometric Langlands correspondence","Howe duality"],"falsifier":"Compute $K^{\\breve{G}}(Z_{\\gamma\\gamma})$ for $G=GL_3$ with $P_\\gamma = B\\{1,s_1\\}B$ (the case in Example 5.10 where no cellular fibration exists). Corollary 5.6 predicts a free $R(\\breve{G})$-module of rank $9$; an independent computation via $\\breve{T}$-localization that yields any other rank falsifies Proposition 5.11 and Theorem 5.21. Alternatively, in the $SO_5$ specialization $Q_f = Q^{\\imath}_2$, the convolution structure constants of $K^{\\breve{G}}(Z_f)$ must match the known presentation of the affine $\\imath$Schur algebra; any mismatch is a falsifier.","tokens_in":47130,"feed_emoji":"🧮","tokens_out":20069,"duration_ms":161341,"temperature":0.7,"pith_summary":"The paper claims that every generalized Schur algebra of any finite type can be built from the geometry of a single kind of variety, the generalized Steinberg variety. For a finite collection of Weyl-group orbits on the weight lattice, the top Borel-Moore homology of that variety, with convolution as multiplication, is the ordinary Schur algebra, while its equivariant K-group is the affine q-Schur algebra. This matters because earlier geometric constructions of Schur algebras existed mainly for type A; the uniform recipe covers all types and yields a Schur-algebra analogue of local geometric Langlands reciprocity, an equivariant K-theoretic realization of quasi-split affine ı-quantum groups of type AIII, and geometric Howe dualities for affine quantum groups.","feed_headline":"Steinberg varieties build Schur algebras of any type","feed_subtitle":"Its top homology gives the Schur algebra; its equivariant K-group gives the affine q-version.","key_machinery":"The central object is the generalized Steinberg variety $Z_f = \\widetilde N_f \\times_N \\widetilde N_f$, whose irreducible components are the conormal bundles $T^*_{O_{\\gamma,w,\\nu}}$ to the $G$-orbits on $F_f \\times F_f$. Convolution in Borel-Moore homology (Theorem 3.11) and in $\\breve{G}$-equivariant $K$-theory (Theorem 5.21) gives the algebra structure. The companion variety $\\widetilde N_f \\times_N \\widetilde N$ carries the Fock-space module $\\widetilde T_f$ for the affine Hecke algebra, and the split injections $\\iota_w$ of Proposition 5.11 supply the $R(\\breve{G})$-basis of the filtration pieces that replaces the unavailable cellular-fibration lemma.","core_discovery":"Theorem 5.21 asserts that for any finite $W$-invariant subset $Q_f$ of the weight lattice, the equivariant $K$-group $K^{\\breve{G}}(Z_f)$ of the generalized Steinberg variety, with its convolution product, is isomorphic as an algebra to the affine $q$-Schur algebra $\\widetilde S_f = \\mathrm{End}_{\\widetilde H}(\\widetilde T_f)$. Theorem 3.11 is the classical-limit statement: the top Borel-Moore homology $H(Z_f)$ is isomorphic to $S_f = \\mathrm{End}_{QW}(T_f)$. The proof avoids the missing cellular-fibration lemma by first identifying the Fock-space module $\\widetilde T_f$ with $K^{\\breve{G}}(\\widetilde N_f \\times_N \\widetilde N)$ as an $\\widetilde H$-module and then transferring the structure to $Z_f$; the $R(\\breve{G})$-basis of each filtration piece is built from the split injections $\\iota_w$ of Proposition 5.11.","pith_inferences":["The module-first strategy, realizing the Fock space before the algebra, should transfer to asymptotic affine q-Schur algebras, where no equivariant K-theoretic description is currently known; the authors mention this as ongoing work.","If the basis in Proposition 5.11 can be made constructive, the main isomorphism would probably yield explicit convolution formulas for the affine $\\imath$-quantum groups of type AIII, including the variants that still lack a Drinfeld-type presentation.","The theorem is stated for groups whose simple factors are simply connected or of type $SO_{2d+1}$; checking whether the same geometric construction works for adjoint exceptional groups would delimit the true scope of 'any type'.","In the three-parameter type-C setting, the same machinery should produce an exotic Steinberg realization of the three-parameter affine $\\imath$-Schur algebras, mirroring the extension from the one-parameter affine Hecke algebra to the exotic type-C algebra."],"forward_implications":["In the classical limit, the basis of $H(Z_f)$ given by fundamental classes of conormal bundles gives a direct geometric presentation of the Schur algebra $S_f$ of any type.","The equivariant $K$-theoretic realization identifies irreducible representations of affine $q$-Schur algebras with data attached to nilpotent orbits and component-group characters, extending the Springer-type classification beyond type A.","Choosing $Q_f$ as a single regular $W$-orbit recovers the known affine Hecke algebra realization, making the main theorem a Schur-algebra analogue of the local geometric Langlands correspondence.","Stabilizing the K-theoretic affine $q$-Schur algebras realizes quasi-split affine $\\imath$-quantum groups of type AIII, including variants for which no explicit Drinfeld presentation is available.","At specializations in type A and type B/C, the double-centralizer property gives geometric Howe dualities between affine quantum groups of type A and between affine $\\imath$-quantum groups of type AIII."],"supporting_citations":[{"why":"It supplies the convolution formalism, the Steinberg-variety machinery, the cellular-fibration lemma, localization, and the affine Hecke algebra realization that the paper uses throughout.","marker":"[CG97]"},{"why":"It introduced the geometric realization of $q$-Schur algebras on partial flag varieties that this paper generalizes from type A to arbitrary type.","marker":"[BLM90]"},{"why":"It defined the generalized $q$-Schur algebras $S_f$ via $W$-orbits and gave their realization as convolution algebras of invariant functions on double flag spaces, whose classical limit Theorem 3.11 reproduces geometrically.","marker":"[LW22]"},{"why":"It introduced the affine $q$-Schur algebras $\\widetilde S_f$ and their realization as convolution algebras of invariant functions on double flag spaces, which Theorem 5.21 realizes through equivariant $K$-theory.","marker":"[CLW24]"},{"why":"It established the equivariant $K$-theoretic realization of affine Hecke algebras that the proof uses to construct the $\\widetilde H$-action on the Fock space.","marker":"[G87]"},{"why":"It provides the freeness and rank computations for representation rings of parabolic subgroups that underpin the rank calculations in Corollary 5.6.","marker":"[St75]"}],"fun_headline_variants":["Steinberg varieties yield Schur algebras in every type","Schur algebras via geometry: Langlands and Howe duality","Borel-Moore homology builds Schur algebras of any type","Equivariant K-theory realizes affine q-Schur algebras","Geometry unites Schur, Langlands, and quantum groups"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that each equivariant $K$-group in the filtration of $Z_{\\gamma\\nu}$ has the claimed $R(\\breve{G})$-basis built from the split injections $\\iota_w$; this is justified in the paper by a one-line induction, in a setting where the standard cellular-fibration lemma is known to fail.","fun_headline_variants_meta":{"raw":{"variants":["Steinberg varieties yield Schur algebras in every type","Schur algebras via geometry: Langlands and Howe duality","Borel-Moore homology builds Schur algebras of any type","Equivariant K-theory realizes affine q-Schur algebras","Geometry unites Schur, Langlands, and quantum groups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000296,"raw_usage":{"total_tokens":1654,"prompt_tokens":819,"completion_tokens":835,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":435,"completion_tokens_details":{"reasoning_tokens":750}},"tokens_in":435,"tokens_out":835,"duration_ms":7935,"temperature":1.0,"reasoning_tokens":750,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:21:35.042212+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $K^{\\breve{G}}(Z_{\\gamma\\gamma})$ for $G=GL_3$ with $P_\\gamma = B\\{1,s_1\\}B$ (the case in Example 5.10 where no cellular fibration exists). Corollary 5.6 predicts a free $R(\\breve{G})$-module of rank $9$; an independent computation via $\\breve{T}$-localization that yields any other rank falsifies Proposition 5.11 and Theorem 5.21. Alternatively, in the $SO_5$ specialization $Q_f = Q^{\\imath}_2$, the convolution structure constants of $K^{\\breve{G}}(Z_f)$ must match the known presentation of the affine $\\imath$Schur algebra; any mismatch is a falsifier.","supporting_citations":[],"review_version":1}