{"id":"fd57af48-4439-4526-ae58-bd7c7922542b","arxiv_id":"2412.01962","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every dominant coweight λ, the global Schubert variety for GL_n equals the lattice-family subset X_λ.","lead":"This paper gives a lattice-theoretic description of global Schubert varieties for GL_n, showing that each such variety coincides with a subset X_λ defined by simple lattice conditions. The result was known from the theory of Shimura varieties; the paper supplies a self-contained proof using linear algebra and topology.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Corollary 6.3 asserts properness of the ad hoc global convolution map m without proof; the induction proving Theorem 1.1 depends on this to conclude the image is closed, so this is the load-bearing gap.","rationale":"I read the paper as a self-contained proof of a known theorem. The theorem itself is supported by prior work, so the central claim is likely true, but the paper's own proof must carry the argument. The weakest point is exactly the global convolution map: the ordinary affine-Grassmannian convolution properness is standard, but the global version is introduced with an ad hoc definition, and the paper even flags that it is ad hoc. Corollary 6.3 then relies on unproved properness of m to conclude closedness of the image, and without closedness the induction proving Theorem 1.1 does not go through. This is a genuine missing proof rather than a disagreement with consensus, and it is localized and likely fixable by invoking the ind-properness of the global affine Grassmannian over C or by proving the corresponding boundedness for the concrete lattice model. The rest of the proof—the minuscule case, the combinatorial Theorem 5.10, and Lemma 6.5—is detailed and plausible, so I do not see reason to move away from the reader's CONDITIONAL verdict. The verdict should remain unchanged.","tokens_in":27858,"tokens_out":12146,"duration_ms":122964,"concrete_test":"For the lattice model of §3, define Gr_{≤M} = { (y,L_1,...,L_n) ∈ Gr | t^{-M}O^n ⊂ L_i ⊂ t^M O^n for all i }. Show that these are closed finite-type pieces exhausting Gr, and for each M find M′ such that m^{-1}(Gr_{≤M}) inside X_λ ~×_C X_ϖ_k is contained in the subset where all L_i and L′_i satisfy the same type of bound; then check that the induced map on each pair of finite-type pieces satisfies the valuative criterion using the explicit maps θ_i(y). If this exhaustion and valuative check go through, Corollary 6.3 is established; if a sequence in the source escapes every Gr_{≤M} while its image is bounded, properness fails and the induction would break.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 1.1 reduces the special-fiber statement to (3.4), and the induction step obtains (3.4) from Corollary 6.3. Corollary 6.3 says that m: X_λ ~×_C X_ϖ_k → Gr has image Gr_{λ+ϖ_k}. The justification is: the generic fiber is C^× × Gr_{λ+ϖ_k} by Lemma 6.1, and 'm is proper' with irreducible domain makes the image closed and irreducible. Irreducibility is proved in Lemma 6.2, but properness is only asserted. Section 6 explicitly defines the global convolution with an 'ad hoc definition' and does not identify m with a map between ind-proper schemes; Section 3 defines Gr as a subset of C × Gr^n and never establishes an ind-proper structure. The ordinary convolution properness used in Lemma 6.1 is standard, but it does not automatically transfer to the global ad hoc space. If m is not proper, its image need not be closed, and the equality with Gr_{λ+ϖ_k} would fail, breaking the induction. This is the least secure step in the self-contained argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies global Schubert varieties for GL_n(C) using a lattice-theoretic model. It defines the global affine Grassmannian Gr as the set of lattice families (y,L_1,...,L_n) satisfying (3.1)-(3.2), and for a dominant coweight λ defines X_λ = Gr ∩ (C × Gr_λ × ... × Gr_λ), where Gr_λ is the ordinary spherical Schubert variety. The main theorem (Theorem 1.1) asserts that X_λ equals the global Schubert variety Gr_λ, i.e. the closure of C^× × Gr_λ. The proof treats fundamental coweights by explicit 1-parameter families in Section 4, then uses a combinatorial statement about alcoves (Theorem 5.10) and a global convolution construction (Section 6) to induct on λ_1 − λ_n.","tokens_in":28106,"tokens_out":9232,"duration_ms":94842,"significance":"The theorem itself is not new; the authors acknowledge that it follows from work of Haines–Ngô and Zhu via nearby cycles and local models. The paper's claimed contribution is a self-contained proof using only topology and linear algebra. If the gaps in Section 6 are repaired, this would be a genuinely useful and accessible account, especially the explicit pluscule-case formulas in Section 4 and the combinatorial Lemma 5.1. The paper is carefully structured and the minuscule case is worked out in substantial detail. However, the proof as written depends on an unproved properness assertion in Corollary 6.3, and on a proof-sketch in Lemma 6.1, so the self-contained claim is not yet fully met.","major_comments":[{"comment":"The corollary asserts without proof that the map m : X_λ ~×_C X_̟_k → Gr is proper. The global convolution space is introduced by an ad hoc definition on page 25, and §3 only gives Gr the structure of an ind-variety as a subset of C × Gr^n; it is not shown to be ind-proper, and because of the C factor it is not. Properness therefore does not follow from the standard properness of ordinary convolution used in Lemma 6.1. The final step of Theorem 1.1 uses this corollary to conclude that L_x ∈ Gr_λ, so without a proof of properness (or a substitute closedness argument) the induction step is incomplete.","section":"§6, Corollary 6.3"},{"comment":"The proof of Lemma 6.1 is only a sketch. The assertion that the largest dominant ν with L_ν in the image of m : Gr_λ ~× Gr_μ → Gr is λ + μ is not proved, and neither is the properness of this ordinary convolution map. These are standard facts in the affine Grassmannian literature, but the paper promises a self-contained proof, and the lemma is used for the generic fiber contribution in Corollary 6.3. Please give a complete proof or a precise reference with all hypotheses verified.","section":"§6, Lemma 6.1"},{"comment":"These lemmas are stated without proof, with only 'we omit their proofs.' They are used in Lemma 5.8 and hence in Theorem 5.10, which is essential for the induction in Theorem 1.1. The rotation formula in Lemma 5.2(3) and the permutation identity in Lemma 5.3 are not entirely immediate from the definitions. Please include proofs, or at least detailed derivations, so that the combinatorial backbone of the induction is fully supported.","section":"§5, Lemmas 5.2 and 5.3"}],"minor_comments":[{"comment":"The notation for the global Schubert variety is overloaded with the ordinary spherical Schubert variety; the parenthetical '(Note that the notation “Gr_λ” is not defined.)' is confusing. Please use distinct symbols (for instance Gr_λ^glob and Gr_λ^sph) or otherwise clarify the typography.","section":"§3, notation after (3.3)"},{"comment":"The remark that the ind-scheme Gr^fancy is not reduced and that the variety (3.3) is its reduced subscheme is helpful, but it should be stated explicitly that all subsequent constructions and the main theorem concern reduced varieties at the level of C-points.","section":"§3, Remark 3.2"},{"comment":"The choice of N 'sufficiently large for all of these lemmas for all pairs (q,i)' is not quantified. This is acceptable, but a short justification that the finitely many conditions can be simultaneously satisfied would improve readability.","section":"§4, Theorem 4.11"},{"comment":"In the description of X ~× Gr_̟_k, the condition tL ⊂ L' ⊂ L is written without specifying which inclusion is strict; the dimension condition dim L'/tL = n−k already enforces the correct quotient, so this is only a minor wording issue.","section":"§6, equation (6.1)"}],"recommendation":"major_revision","confidential_remarks":"The central theorem is not new, and the paper's value lies in its self-contained elementary proof. The referee report focuses on whether that proof is complete. The properness gap in Corollary 6.3 is serious because the induction for Theorem 1.1 depends on the image being closed. If the authors can prove properness directly or replace that step with a closedness argument, and fill the proof sketches in Lemmas 6.1 and 5.2–5.3, the paper would be suitable for publication. The pluscule case appears sound and is a strong part of the paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thesis of this paper is not new — Theorem 1.1 can be deduced from Haines–Ngô or Kottwitz–Rapoport plus Zhu, and the authors say so in the introduction. What they add is a self-contained lattice-theoretic proof for GL_n, with the minuscule case done by explicit limit formulas. That is a legitimate contribution, mostly expository but with real content in Section 4.\n\nWhat the paper does well: the explicit formulas in Section 4 are intricate and look correct — they genuinely exhibit every special-fiber point as a limit of a 1-parameter family. The combinatorial Section 5 is technical but seems to deliver exactly what the convolution induction needs. The introduction is candid about prior work, which I appreciate.\n\nThe soft spots are concentrated in Section 6. The stress-test is right: Corollary 6.3 asserts that the global convolution map m : X_λ ~×_C X_̟_k → Gr is proper, and gives no proof. This is the step that makes the image closed and irreducible, so the induction in Theorem 1.1 depends on it. The assertion is likely true — the global affine Grassmannian is ind-proper over C, and the convolution space is a closed subvariety of the fiber product — but the paper never establishes that, and the ad hoc definition does not make it automatic. Lemma 6.1 is also only a proof sketch, and Lemmas 5.2–5.3 are stated without proof. These are fixable gaps, but they do undercut the claim of a fully self-contained argument.\n\nFor whom: this will be most useful to people who want an elementary route to a known result, or who teach local models and want a bridge from the Shimura-variety literature to concrete lattice statements. It is not a new-result paper, so the value is pedagogical and structural.\n\nRecommendation: yes, send it to peer review. Ask the authors to prove properness of the global convolution map (or cite a standard reference for ind-properness of Gr over C), expand Lemma 6.1, and either prove or explicitly defer Lemmas 5.2–5.3. After those revisions I would be comfortable with it.","headline":"Known theorem, honest introduction, useful self-contained proof with a fixable properness gap in the convolution step.","tokens_in":28636,"tokens_out":4964,"would_cite":false,"duration_ms":255819,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14M15","14G35"],"pacs":[],"model":"deepseek-v4-flash","headline":"For GL_n, every global Schubert variety is exactly the family of lattices lying in an ordinary Schubert variety.","keywords":["global Schubert variety","affine Grassmannian","lattice families","dominant coweight","Kottwitz-Rapoport alcoves","convolution spaces","local models of Shimura varieties","GL_n"],"falsifier":"For a specific coweight such as λ = (2,0) for GL_2, compute X_λ ∩ F_l directly from the lattice inequalities of Remark 3.5 and compare it with the union of I-orbits in the closure of C^× × Gr_λ; if any λ-permissible alcove x gives a lattice chain L_x outside that closure, the equality Gr_λ = X_λ would be false.","tokens_in":27668,"feed_emoji":"📐","tokens_out":6563,"duration_ms":58315,"temperature":0.7,"pith_summary":"This paper proves that for the general linear group GL_n, every global Schubert variety is completely described by a simple lattice condition: a point of the global affine Grassmannian lies in the global Schubert variety Gr_λ exactly when all its lattices lie in the ordinary spherical Schubert variety Gr_λ. The authors give a self-contained proof using only basic topology and linear algebra, avoiding Shimura varieties and nearby cycles. The result matters because it is the lattice-theoretic form of the 'topological flatness' of certain local models of Shimura varieties at the level of complex Laurent series. The proof proceeds through an explicit formula for fundamental coweights and an induction using global convolution spaces.","feed_headline":"Lattice families capture all GL_n Schubert varieties","feed_subtitle":"A self-contained proof shows the closure of C× × Gr_λ equals the family of lattices inside Gr_λ.","key_machinery":"The central object is the global affine Grassmannian Gr, realized as the set of lattice families (y, L_1, . . . , L_n) in C × Gr^n satisfying valuation and containment conditions, with special fiber the affine flag variety F_l and generic fiber C^× × Gr. The paper defines X_λ as the closed sub-ind-variety Gr ∩ (C × Gr_λ × · · · × Gr_λ) and proves Gr_λ = X_λ. The proof is carried by two mechanisms: explicit vector and matrix formulas for the fundamental coweight case that produce one-parameter degenerations to any desired point of the special fiber, and the global twisted convolution space X_λ ~×_C X_{̟_k} equipped with a morphism m to Gr whose image is shown to be Gr_{λ+̟_k}. A key auxiliary result is Theorem 5.10, which constructs, for λ = μ + ̟_t, a μ-permissible alcove y such that a given λ-permissible alcove x is in relative position ̟_t with respect to y, enabling the induction.","core_discovery":"On its own terms, the paper's central claim is Theorem 1.1: for any dominant coweight λ, the global Schubert variety Gr_λ, defined as the closure of C^× × Gr_λ in the global affine Grassmannian Gr, equals X_λ, the closed subset of Gr consisting of lattice families (y, L_1, . . . , L_n) with every L_i in Gr_λ. Since X_λ is visibly closed and agrees with Gr_λ over C^×, the whole content is that its special fiber over 0 is not too large: every I-orbit in X_λ ∩ F_l is actually in the closure. The authors establish this first for fundamental coweights, where each point of the special fiber is exhibited as the limit of an explicit one-parameter family, and then for general λ by induction, using a combinatorial statement about Kottwitz–Rapoport alcoves and the geometry of global convolution spaces.","pith_inferences":["The ad hoc construction of the global convolution space in Section 6 suggests a natural check: proving properness of m directly would let the same induction work without the asserted closed-image property, and would likely transfer the argument to reductive groups admitting a lattice model.","The explicit families in Section 4 could be composed with the alcove rotations of Section 5 to produce explicit degenerations for arbitrary dominant coweights, not just fundamental ones.","In equal characteristic, Theorem 1.1 is the lattice-theoretic form of topological flatness of local models; carrying the same lattice-family language into mixed characteristic could give an elementary path to the corresponding flatness statements."],"forward_implications":["Membership in a global Schubert variety for GL_n becomes checkable by explicit lattice inequalities: a tuple (y, L_1, . . . , L_n) is in Gr_λ exactly when every L_i satisfies the valuation and dimension bounds of Remark 2.1.","The special fiber of Gr_λ over 0 is the union of the I-orbits labelled by the λ-permissible alcoves, matching the description Zhu proved using admissible alcoves.","For minuscule coweights, each point of the special fiber is obtained as the limit of a one-parameter family running through the generic part C^× × Gr_λ.","The result gives an elementary proof, using only basic topology and linear algebra, of the topological-flatness statement for these GL_n local models, without Shimura varieties or nearby cycles."],"supporting_citations":[{"why":"Introduces a version of the global affine Grassmannian whose generic and special fibers this paper adopts as its starting point.","marker":"[3]"},{"why":"Proved the minuscule case of Theorem 1.1 by explicit calculation, the same case treated here with different formulas.","marker":"[4]"},{"why":"Introduced λ-permissible alcoves and the alcove combinatorics that Section 5 and Theorem 5.10 rely on.","marker":"[9]"},{"why":"Supplies the parahoric local-model formalism and the 'fancy' global affine Grassmannian to which the lattice-family description is compared.","marker":"[10]"},{"why":"Gives the ind-scheme construction of the global affine Grassmannian and the special-fiber description via admissible alcoves that motivates the main result.","marker":"[13]"},{"why":"Originated lattice descriptions of local models of Shimura varieties, the circle of ideas that Theorem 1.1 recovers in equal characteristic.","marker":"[11]"}],"fun_headline_variants":["Lattice families equal GL_n global Schubert varieties","GL_n Schubert closures are exactly lattice families","Theorem: Schubert variety equals lattice-set closure","Global Schubert variety described by lattice conditions","Explicit limits prove lattice model for GL_n Schubert"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that the map m from the global convolution space X_λ ~×_C X_{̟_k} to the global affine Grassmannian is proper, and therefore has closed image; this properness is asserted in Corollary 6.3 but not proved in the paper.","fun_headline_variants_meta":{"raw":{"variants":["Lattice families equal GL_n global Schubert varieties","GL_n Schubert closures are exactly lattice families","Theorem: Schubert variety equals lattice-set closure","Global Schubert variety described by lattice conditions","Explicit limits prove lattice model for GL_n Schubert"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000176,"raw_usage":{"total_tokens":1190,"prompt_tokens":745,"completion_tokens":445,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":361,"completion_tokens_details":{"reasoning_tokens":372}},"tokens_in":361,"tokens_out":445,"duration_ms":4714,"temperature":1.0,"reasoning_tokens":372,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T23:58:20.599979+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a specific coweight such as λ = (2,0) for GL_2, compute X_λ ∩ F_l directly from the lattice inequalities of Remark 3.5 and compare it with the union of I-orbits in the closure of C^× × Gr_λ; if any λ-permissible alcove x gives a lattice chain L_x outside that closure, the equality Gr_λ = X_λ would be false.","supporting_citations":[{"cited_title":"Gaitsgory, Construction of central elements in the affine Hecke algebra v ia nearby cycles , Invent","cited_arxiv_id":null,"evidence_quote":"Introduces a version of the global affine Grassmannian whose generic and special fibers this paper adopts as its starting point."},{"cited_title":"G¨ ortz, On the flatness of certain models of Shimura varieties of PEL- type, Math","cited_arxiv_id":null,"evidence_quote":"Proved the minuscule case of Theorem 1.1 by explicit calculation, the same case treated here with different formulas."},{"cited_title":"Kottwitz and M","cited_arxiv_id":null,"evidence_quote":"Introduced λ-permissible alcoves and the alcove combinatorics that Section 5 and Theorem 5.10 rely on."},{"cited_title":"Pappas and X","cited_arxiv_id":null,"evidence_quote":"Supplies the parahoric local-model formalism and the 'fancy' global affine Grassmannian to which the lattice-family description is compared."},{"cited_title":"Zhu, On the coherence conjecture of Pappas and Rapoport , Ann","cited_arxiv_id":null,"evidence_quote":"Gives the ind-scheme construction of the global affine Grassmannian and the special-fiber description via admissible alcoves that motivates the main result."},{"cited_title":"Rapoport and T","cited_arxiv_id":null,"evidence_quote":"Originated lattice descriptions of local models of Shimura varieties, the circle of ideas that Theorem 1.1 recovers in equal characteristic."}],"review_version":1}