{"id":"fe2062fc-9cd0-4f96-8023-a93f61e2f7b3","arxiv_id":"2505.09905","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A canonical orbifold resolution of any Richardson variety is constructed from equivariant stable map spaces; its boundary dual complex is the order complex of an open Bruhat interval, and in the Grassmannian case its T-fixed points are indexed by rim-hook tableaux.","lead":"This paper builds a canonical, choice-free resolution of singularities for Richardson varieties (intersections of Schubert varieties) as smooth orbifolds using spaces of stable maps. The construction connects a classical combinatorial object, the Bruhat interval, to the topology of these resolutions and gives explicit Betti number formulas in the Grassmannian and flag cases.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"General-G/P boundary claim likely fails: equivariantly smoothable caterpillars force a single S-weight, so Bruhat chains mixing different rho-check weights are empty; Theorem 4's proof works only because type A has constant weight.","rationale":"The reader's stated weakest assumption, Lemma 1(2), is less serious than it appears: for a closed T-invariant subvariety X of G/P, the two Bialynicki-Birula decompositions of X are induced from the Bruhat stratifications of G/P, so the stratification hypothesis holds automatically. That part of Theorem 2 is not the bottleneck. The real load-bearing issue is the step from type A to general G/P. The dual-complex/sncd assertion is the combinatorial heart of the main claim, and its proof (Theorem 4(1)-(2)) explicitly depends on equivariant smoothability of sutured caterpillars. Corollary 1 forces one common S-weight on any equivariantly smoothable caterpillar, and this is exactly what makes suturing work in type A. In most other G/P the rho-check weights of T-fixed curves vary, so chains mixing different weights cannot be smoothed equivariantly; the corresponding strata are empty and the dual complex is not the full order complex. The paper proves Theorem 4 only in the cases where the obstruction is vacuous, yet the abstract announces the result for all generalized flag manifolds. The type-A results, the GKM weights, and the rim-hook-tableau Betti number computations are substantial and should stand, but the general G/P boundary claim needs either a proof of the weight-constraint or a revised statement. CONDITIONAL is therefore the appropriate verdict: accept the type-A core while requiring the general claim to be corrected or carefully qualified.","tokens_in":22826,"tokens_out":28670,"duration_ms":337577,"concrete_test":"Take G=Sp_4, P=B, with rho-check the smallest regular dominant coweight, so the coordinate weights in the standard 4-dimensional representation are (3/2, 1/2, -1/2, -3/2). Choose u=1 and v a Bruhat element containing both a short-root reflection such as s_{e1-e2} (S-weight 1) and a long-root reflection such as s_{2e1} (S-weight 3) in an increasing chain. Compute the T-fixed points of X_v^u//StMap^rho-check by applying the reflection-chain recipe of Section 5.2 with the type-C Weyl group action. If the mixed-weight chain is not realized by any fixed point, while the full order complex of the interval (u,v) would require it, then Theorem 4(2) and the claimed dual-complex statement fail outside type A.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Corollary 1 classifies every fixed point of X//StMap^S as a caterpillar all of whose components carry one common S-weight alpha (equivalently, all components are eT-isomorphic). Hence a stratum indexed by a Bruhat chain kappa_0<...<kappa_{m+1} can meet the quotient only if the factors X_{kappa_i}^{kappa_{i+1}}//StMap can be chosen with that same alpha. In the cases where Theorem 4 is actually proved (Grassmannians and GL_n flag manifolds) this condition is automatic: for rho-check, every T-fixed curve has the same S-weight because the coordinate weights form an arithmetic progression. For a general G/P the rho-check-pairings of roots take several values; for example in Sp_4/B, short-root steps have weight 1 and long-root steps have weights 2 or 3. If a chain mixes two such steps, the node-deformation summand Tn^- tensor Tn^+ has nonzero S-weight, so Proposition 2's duality fails and no S-equivariant smoothing exists; the sutured curve lies in a different component of the S-fixed locus and is not in X//StMap^S. Thus the identification of the dual complex with the full Bruhat order complex, which is the part of the central claim covering all G/P, is not merely unproved but likely false as stated. The type-A restriction in Theorem 4 is essential, and the abstract overstates the general result.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces, for a projective variety M with a C^*-action S, a 'stable map quotient' M//StMap^S, defined as the connected component of the S-fixed locus of a Fulton--Pandharipande stable map space containing generic S-orbit closures. For a Richardson variety X_λ^μ in a flag variety G/P, the author defines a resolution fX_λ^μ via a graph-space variant of this quotient and claims that it is a canonical, choice-free orbifold resolution. Further claims are that the boundary is an orbifold simple normal crossings divisor whose dual simplicial complex is the order complex of the open Bruhat interval (λ,μ), and that in the Grassmannian case the T-fixed points are indexed by standard rim-hook tableaux with explicit GKM weights and computable Betti numbers. The paper contains detailed combinatorial computations in type A and a discussion, partly conjectural, of anticanonicality of the boundary.","tokens_in":23108,"tokens_out":6749,"duration_ms":73252,"significance":"If the type-A results are correct, the paper gives a genuinely new and useful construction: a choice-free birational orbifold model of Richardson varieties with computable equivariant cohomology, connecting stable map spaces to Chow quotients, Bruhat order, and Deodhar-style decompositions. The rim-hook-tableaux indexing, the explicit GKM weights in §5--§6, and the concrete Betti-number examples are valuable and appear to be carefully computed. The relation to the Björner--Wachs theorem is elegant and, in the cases where Theorem 4 is proved, gives a geometric realization of the order complex of an open Bruhat interval. However, the advertised general-G/P statement is not supported by the proofs and, on the basis of the paper's own Corollary 1, appears to be false in that generality. The significance of the paper therefore depends on a substantial reframing: the main theorems must be restricted to the cases actually proved, and the general-G/P claims must be either proved or removed.","major_comments":[{"comment":"The abstract and Theorem 5 state the dual-complex/order-complex result without a type restriction, but Theorem 4 is proved only under the explicit assumption 'G/P a Grassmannian or full flag variety.' This restriction is not cosmetic. Corollary 1 forces every point of M//StMap^S to be a caterpillar all of whose components carry one common S-weight α. Consequently, a stratum indexed by a Bruhat chain κ_0<...<κ_{m+1} can meet the quotient only if all factors X_{κ_i}^{κ_{i+1}}//StMap^S are realizable with the same α. For ρ̌ on a general G/P, the root-pairing weights take several values; for example, in Sp_4/B a short-root step has weight 1 and long-root steps have weights 2 or 3. A chain mixing such steps has no S-equivariant smoothing, because the node-deformation summand T_n^-⊗T_n^+ has nonzero S-weight. The identification of the dual complex with the full order complex of (λ,μ) is therefore not merely unproved in general but false as stated. Theorem 5 must be restricted to the type-A cases, and the abstract must be corrected accordingly.","section":"Abstract, §1, and §4 (Theorem 4, Theorem 5)"},{"comment":"Theorem 2's proof uses Lemma 1(2) to conclude that the image of a fixed stable map lies in X_λ ∩ X^μ. Lemma 1(2) is conditional on the hypothesis that the two Białynicki-Birula decompositions of X_λ^μ defined by ρ̌ are stratifications, and the text says this hypothesis is 'likely unnecessary' but supplies no proof. For arbitrary G/P, Richardson varieties can be singular, and the Białynicki-Birula decompositions are not verified to be stratifications. Since Theorem 2 is the basis for the orbifold-smoothness of fX_λ^μ for all G/P, this is a load-bearing gap. The paper should either prove the stratification hypothesis for Richardson varieties or explicitly state Theorem 2 and the resolution theorem under this hypothesis.","section":"§2.2 (Lemma 1) and §2.4 (proof of Theorem 2)"},{"comment":"The suturing argument in Theorem 4(1) requires that the curve obtained by gluing caterpillars along fixed points is equivariantly smoothable, i.e. lies in the same connected component of the S-fixed stable map space. The text says 'We'll need a converse to proposition 2' and then asserts that convexity of G/P plus the proof of Theorem 2 provides it. No actual deformation argument is given. For general G/P the asserted converse is false when adjacent components have different S-weights, since the corresponding node-deformation summand T_n^-⊗T_n^+ has nonzero S-weight and cannot be smoothed equivariantly. This is exactly the mechanism that makes the type-A restriction in Theorem 4 essential. The proof needs either a genuine deformation-theoretic argument in the cases where the statement is claimed, or a revised statement limited to the cases where the weight condition holds.","section":"§4, proof of Theorem 4(1) and §4.2"}],"minor_comments":[{"comment":"The first sentence of the abstract says 'generalized partial flag manifold' and the dual-complex sentence carries no type restriction; please add the Grassmannian/full-flag hypothesis or explicitly mark the general-G/P cases as open.","section":"Abstract and §1"},{"comment":"The notation (M_μ^λ//StMap^S)_{k nodes} uses subscripts that are not defined in the statement; it should be M//StMap^S or the subscripts should be introduced.","section":"§2.8, Proposition 3"},{"comment":"The phrase 'a [pt/Z2] which is not a stratum' is informal; please specify the orbifold chart and how the Z_2 stabilizer arises at that T-fixed point.","section":"§8.3, Proposition 6"},{"comment":"The entry [KaStZe91] is missing the first author's name; it should read 'M. M. Kapranov, B. Sturmfels, and A. V. Zelevinsky.'","section":"References"},{"comment":"The sentence 'Michalek and Wang have informed me...' would read better as a footnote or an acknowledgment rather than as part of the mathematical narrative.","section":"§3"}],"recommendation":"major_revision","confidential_remarks":"The type-A content of the paper is substantial and likely correct, and the combinatorial computations are a genuine contribution. My main concern is framing: the abstract and Theorem 5 present the dual-complex statement as a general theorem although Theorem 4 is proved only for Grassmannians and full flags, and the internal logic of Corollary 1 indicates that the general-G/P statement is false, not merely unproved. If the authors are willing to restrict the main claims to the cases actually established and to address the Lemma 1(2) hypothesis, the paper is a strong contribution. If they insist on retaining the unrestricted general-G/P claims, I would not recommend publication. The relation to the announced work of Michalek, Monin, and Wang should also be stated more precisely once that work is available."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real construction—stable map quotients give a choice-free orbifold resolution of Richardson varieties, with careful GKM computations in type A and a clean relation to Chow quotients. Read it for that. But the abstract promises the dual complex result for all G/P, and only Grassmannian and full-flag cases are proved—and I think the general claim is not just unproved but wrong.\n\nWhat's new and good: M//StMap^S is a sensible quotient, the graph-space trick to restore dimension is standard but well executed, and the paper delivers explicit weights, rim-hook tableau indexing of T-fixed points, Betti numbers, and GKM graphs for Grassmannians. The author is honest about the wobbly bits—Lemma 1(2)'s stratification hypothesis is flagged and its removal left open. That's fine if the main results don't depend on it, but Theorem 2 does at present.\n\nThe soft spot that matters is bigger. Corollary 1 says every equivariantly smoothable fixed curve is a caterpillar whose components all carry the same S-weight. When you suture components indexed by a Bruhat chain κ0<...<κm+1, the node-deformation summand Tn^-⊗Tn^+ must have trivial S-weight; otherwise Proposition 2's duality fails and the sutured curve is not in the stable map quotient. In type A with ρ̌, every T-fixed curve has normalized S-weight 1, so any chain works. In a general G/P, ρ̌ pairs differently with short and long roots—e.g. Sp4/B has steps with weights 1,2,3. A chain mixing such steps would have no S-equivariant smoothing, so the stratum is empty. Hence the dual complex is not the full Bruhat order complex for general G/P; the type A restriction in Theorem 4 is essential and the abstract overstates the result.\n\nOther soft spots are minor: the line-bundle section is conditional on orbifold GKM statements the author couldn't find in the literature, and quotient stack vs coarse space issues are handled a bit quickly. But the computations are concrete and checkable, and the combinatorial payoffs (palindromic Betti numbers, Deodhar-style decomposition) are genuinely interesting.\n\nMy take: this deserves a serious referee. The type A content is strong enough to survive a correction to the general statement. The referee should ask the author to either prove the general dual-complex claim under whatever extra hypothesis is really needed, or—more likely—rewrite the abstract and Theorem 4 to say what is actually proved.","headline":"A genuinely new canonical orbifold resolution of Richardson varieties with strong type A results; the general G/P dual-complex claim in the abstract looks false as stated.","tokens_in":23649,"tokens_out":3853,"would_cite":true,"duration_ms":40871,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14M15","14N35","14D23"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every Richardson variety in a generalized partial flag manifold has a canonical orbifold resolution, built from equivariant stable maps, whose boundary strata are indexed by the open Bruhat interval.","keywords":["Richardson varieties","stable maps","orbifold resolutions","simple normal crossings divisors","Bruhat order","dual simplicial complex","GKM spaces","rim-hook tableaux"],"falsifier":"Compute, for a small singular Richardson variety in a flag manifold where the two one-sided limit decompositions are not known to be stratifications, the full $\\check{\\rho}$-fixed locus in the graph space $X_{(\\lambda,0)}^{(\\mu,\\infty)}/\\!/\\,\\mathrm{StMap}^{\\check{\\rho}_\\Delta}$; if any fixed caterpillar in the chosen component has an end mapping outside $X_\\lambda\\cap X^\\mu$, Theorem 2's component identification fails and the resolution claim for that $X_\\lambda^\\mu$ collapses. In the Grassmannian case, a faster check is to compare the Betti numbers predicted by the rim-hook-tableau counts (for instance $1,28,235,787,1167,787,235,28,1$ for $X_{(2)}^{(4,4,2)}$) with the actual rational cohomology of the coarse Chow quotient.","tokens_in":22604,"feed_emoji":"📐","tokens_out":13027,"duration_ms":116882,"temperature":0.7,"pith_summary":"The paper aims to show that every Richardson variety $X_\\lambda^\\mu = X_\\lambda \\cap X^\\mu$ in a generalized partial flag manifold $G/P$ has a canonical resolution of singularities, built from equivariant genus-zero stable maps, and that the construction is free of choices such as reduced words. The resolution $\\tilde{X}_\\lambda^\\mu$ is a smooth orbifold (Deligne-Mumford stack) rather than a manifold, so this is an orbifold log resolution. Its boundary, formed by nodal curves, is an orbifold simple normal crossings divisor, and the paper computes the dual simplicial complex of that divisor: it is the order complex of the open Bruhat interval $(\\lambda,\\mu)$, known by earlier shellability results to be a sphere or a ball. In the Grassmannian case the resolved space is a GKM space whose torus-fixed points are standard rim-hook skew tableaux, and the computed isotropy weights yield rational Betti numbers. The payoff is a canonical combinatorial resolution that connects singularities of Richardson varieties to the topology of Bruhat intervals.","feed_headline":"Stable map quotients give Richardson varieties canonical resolutions","feed_subtitle":"The boundary's dual simplicial complex is the Bruhat interval's order complex, a sphere or ball.","key_machinery":"The central construction is the stable map quotient $M/\\!/\\,\\mathrm{StMap}^S$: for a projective variety $M$ with a circle action $S$, take the connected component of the $S$-fixed locus of the moduli stack of genus-zero stable maps in the class of a generic $S$-orbit closure that contains the maps $z\\mapsto S(z)\\cdot u$, $u$ general. A key structural fact is that every fixed curve in this component is a `caterpillar': a chain of $\\mathbb{P}^1$s, each component carrying the same pair of isotropy weights $\\pm\\alpha$ for a finite extension $\\tilde{S}$ of $S$, with no collapsed components. The dimension loss from this quotient is repaired by passing to the graph space $(M\\times\\mathbb{P}^1)/\\!/\\,\\mathrm{StMap}^{S_\\Delta}$, in which exactly one component maps isomorphically to $\\mathbb{P}^1$, so evaluation at $1$ gives the birational morphism to $M$. The tangent-space analysis splits deformations into node-gluing summands and invariant sections of $\\gamma^*TM$; the paper computes these $\\tilde{S}$-weights explicitly in the Grassmannian and full-flag cases, yielding the stratification by nodes, the simple-normal-crossings property, the order complex of the Bruhat interval as dual complex, and the GKM/Betti number data.","core_discovery":"On the paper's own terms, the discovery is that the stable map quotient $$\\tilde{X}_\\$\\lambda$^\\mu \\;:=\\; X_{(\\$\\lambda$,0)}^{(\\mu,\\infty)}\\,/\\!/\\,\\mathrm{StMap}^{\\check{\\rho}_\\$\\Delta$}$$ is a smooth Deligne-Mumford stack equipped with a birational morphism to $X_\\lambda^\\mu$, hence a canonical orbifold resolution of the Richardson variety. Here $X_{(\\lambda,0)}^{(\\mu,\\infty)}$ is the Richardson variety in $(G\\times SL_2)/(P\\times B)$ obtained by placing the two Schubert conditions at the $0$- and $\\infty$-sections of $\\mathbb{P}^1$, and $\\check{\\rho}_\\Delta$ is the smallest regular dominant coweight of $G\\times SL_2$. The main mechanism is a component-identification theorem: inside the $\\check{\\rho}$-fixed locus of the stable map space for the generic orbit-closure class, the connected component containing the graph construction is shown, via a lemma on endpoints of fixed caterpillars, to consist exactly of stable maps whose images lie in $X_\\lambda\\cap X^\\mu$. The paper also proves that the boundary of $\\tilde{X}_\\lambda^\\mu$ is a simple normal crossings divisor on the orbifold, with dual simplicial complex equal to the order complex of the open Bruhat interval $(\\lambda,\\mu)$, and develops the tangent-weight computations that make the Grassmannian case a GKM space.","pith_inferences":["An extension the paper leaves open: if the stratification hypothesis in Lemma 1(2) is removable as suspected, the canonical resolution should exist for all Richardson varieties without extra hypotheses, so the smoothness theorem would hold in broader generality than the proof currently shows.","The ball-versus-sphere behavior of the dual complex, together with the orbifold stabilizers found on one-dimensional strata, suggests a refinement of the folk conjecture on dual complexes of anticanonical sncds: boundary components meeting orbifold strata should correspond to the boundary of a ball, not to a sphere.","Once the orbifold analogue of equivariant K-theoretic localization is available, the function $\\Phi$ constructed in Proposition 5 should define a genuine $T$-equivariant ample orbifold line bundle, making $\\tilde{X}_\\lambda^\\mu$ projective in the Grassmannian case; the GKM check in the paper is exactly the condition such a bundle would need to satisfy.","Because the resolution uses no choices beyond the canonical coweight $\\check{\\rho}$, it is a natural candidate for a functorial resolution under the projection maps $\\pi_k:\\mathrm{Fl}(n)\\to \\mathrm{Gr}(k,n)$ that the paper already uses in Section 7.2."],"forward_implications":["Every Richardson variety $X_\\lambda^\\mu\\subseteq G/P$ acquires a canonical resolution by a smooth orbifold, with no reduced word or auxiliary choice entering the construction.","The boundary of $\\tilde{X}_\\lambda^\\mu$ is an orbifold simple normal crossings divisor, so the resolution is a log resolution; its dual simplicial complex is the order complex of the open Bruhat interval $(\\lambda,\\mu)$.","For $G/B$ the dual complex is a sphere, while for general $G/P$ it can be a sphere or a ball, exactly following the known sphere-versus-ball shellability dichotomy for Bruhat order.","For Grassmannian Richardson varieties, $\\tilde{X}_\\lambda^\\mu$ is a GKM space, its $T$-fixed points are standard rim-hook skew tableaux of shape $\\mu\\setminus\\lambda$, and the computed weights give palindromic rational Betti numbers.","The coarse moduli space of $\\tilde{X}_\\lambda^\\mu$ is the Chow quotient $X_\\lambda^\\mu/\\!/_{\\mathrm{Chow}}\\check{\\rho}$, and on the equivariantly smoothable locus the cycle map is bijective."],"supporting_citations":[{"why":"Supplies the smooth-orbifold structure of genus-zero stable map spaces for flag manifolds (Theorem 1), the base of the whole construction.","marker":"[FuPan97]"},{"why":"Provides the proof that the $\\tilde{S}$-weights on the two tangent directions at a node are dual, which forces fixed curves to be caterpillars.","marker":"[CoDesNSuWe22]"},{"why":"Gives the three-term tangent-space formula for stable maps used to prove the stratification, the sncd property, and the Betti-weight computations.","marker":"[Kw07]"},{"why":"Establishes that the order complex of an open Bruhat interval is shellable, hence a sphere or ball; Theorem 5 rests on this.","marker":"[BjWa82]"},{"why":"Defines the dual simplicial complex of a simple normal crossings divisor and states the folklore conjecture relating anticanonical boundaries to spheres, which the paper tests in the orbifold setting.","marker":"[KoX16]"},{"why":"Provides the companion description of dual complexes and weight filtrations used in Section 4.","marker":"[Pay13]"}],"fun_headline_variants":["Stable map quotients give canonical orbifold resolutions","Richardson varieties get orbifold resolutions via stable maps","Bruhat order complex appears as boundary of Richardson resolution","Stable map quotients resolve Richardson varieties, yielding Bruhat order complex","Grassmannian case gives GKM space indexed by rim-hook tableaux"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof needs the two one-sided limits of the circle action on the Richardson variety to decompose the variety into well-behaved cells in a compatible way; for singular Richardson varieties in general flag manifolds this is not automatic, and the paper says the assumption is probably removable but does not prove it.","fun_headline_variants_meta":{"raw":{"variants":["Stable map quotients give canonical orbifold resolutions","Richardson varieties get orbifold resolutions via stable maps","Bruhat order complex appears as boundary of Richardson resolution","Stable map quotients resolve Richardson varieties, yielding Bruhat order complex","Grassmannian case gives GKM space indexed by rim-hook tableaux"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000923,"raw_usage":{"total_tokens":3991,"prompt_tokens":1014,"completion_tokens":2977,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":630,"completion_tokens_details":{"reasoning_tokens":2888}},"tokens_in":630,"tokens_out":2977,"duration_ms":20045,"temperature":1.0,"reasoning_tokens":2888,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:22:39.854200+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, for a small singular Richardson variety in a flag manifold where the two one-sided limit decompositions are not known to be stratifications, the full $\\check{\\rho}$-fixed locus in the graph space $X_{(\\lambda,0)}^{(\\mu,\\infty)}/\\!/\\,\\mathrm{StMap}^{\\check{\\rho}_\\Delta}$; if any fixed caterpillar in the chosen component has an end mapping outside $X_\\lambda\\cap X^\\mu$, Theorem 2's component identification fails and the resolution claim for that $X_\\lambda^\\mu$ collapses. In the Grassmannian case, a faster check is to compare the Betti numbers predicted by the rim-hook-tableau counts (for instance $1,28,235,787,1167,787,235,28,1$ for $X_{(2)}^{(4,4,2)}$) with the actual rational cohomology of the coarse Chow quotient.","supporting_citations":[],"review_version":1}