{"id":"6c9704a0-c196-4605-865a-528ec10cd5fb","arxiv_id":"2501.09962","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Gluable maps of reductive groups make Coulomb branches compose via Hamiltonian reduction, yielding a proof that T^*(G/U_P) for GL_n and SL_n is a Coulomb branch.","lead":"The paper proves a gluing rule for Coulomb branches, spaces from 3d gauge theory, and uses it to show that the affine closure of the cotangent bundle of a parabolic base affine space for GL_n or SL_n is a Coulomb branch. This confirms a physics-inspired conjecture and gives a new structural tool for computing such spaces.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Gluability is not verified for the map GLQm/Gm → (GLA_m × GLAn)/Gm used to prove Theorem 3.3, leaving the proof of Theorem 1.7 conditional on an unstated check.","rationale":"The reader identified the gluable hypothesis as the weakest assumption, and I agree that it is the main technical soft spot. However, the sharper concern is not merely that the hypothesis may be unnecessary in general; it is that the paper's proof of the headline theorem invokes Theorem 1.1 for a specific map without carrying out the finite verification that the hypothesis holds. This is load-bearing because Theorem 3.3 is the bridge from the functoriality theorem to the Bourget-Dancer-Grimminger-Hanany-Zhong conjecture. If the map fails gluability, the proof does not establish the conjecture, even if the conjecture itself is true and could be reached by other means. The check I propose is concrete and would settle whether the gap is real. I found no evidence of internal inconsistency, circular reasoning, or data fitting; the paper is careful and the main construction is plausible. The missing check is exactly the kind of verification a conditional acceptance should demand, so I do not change the reader's verdict.","tokens_in":26082,"tokens_out":19495,"duration_ms":202496,"concrete_test":"For the partition n = 3 = 2+1 (and, if time permits, n = 4 = 3+1), enumerate all weights of N_Qm on a maximal torus of GL_Qm/Gm and their images under the map to (GL_A_m × GL_An)/Gm. For every pair of weights, check whether condition (1) of Definition 2.6 holds with α ∈ Q and, if so, whether some cocharacter of the source torus gives opposite signs as in condition (2). If any violating pair exists, the proof of Theorem 3.3 has a genuine gap; if none exists, the missing verification is supplied and the application of Theorem 1.1 is justified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline application rests on Theorem 3.3, whose proof applies Theorem 1.1 to the map GLQ_m/Gm → (GLA_m × GLAn)/Gm and immediately obtains isomorphism (3.5). The gluable condition of Definition 2.6 is not verified for this map anywhere in Section 3.3. Lemma 3.1 verifies gluability for quiver dismemberments, but the map used here is a map to a product of two quiver gauge groups, not obviously a dismemberment covered by that lemma. Because Remark 1.3 concedes that the gluable hypothesis is likely unnecessary and plans future work, the proof may be tacitly relying on that unproven generalization. If some pair of weights of N_Qm satisfies Definition 2.6(1) with α < 0 and has a cocharacter with opposite signs, Theorem 1.1 does not apply and the identification T^*(G/U_P) ≅ M_C(GL_Qm, N_Qm) has no support from the paper's main theorem. The missing verification is finite and explicit: the weights of the quiver representation are coordinate characters, so Definition 2.6 can be checked by a small computation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a functoriality theorem for Coulomb branches of 3d N=4 gauge theories: for a map H→G of complex reductive groups and a representation N satisfying a hypothesis called gluable (Definition 2.6), the Coulomb branch M_C(H,N) is canonically isomorphic to the balanced product M_C(H,0) ×^{M_C(G,0)} M_C(G,N) over the invariant adjoint quotient h//H (Theorem 1.1). The theorem is applied to show that Coulomb branches of loopless quivers are determined by two-vertex quivers (Corollary 1.5) and to prove the Bourget–Dancer–Grimminger–Hanany–Zhong conjecture for GL_n and SL_n: the affine closure of T*(G/U_P) is isomorphic to an explicit quiver Coulomb branch (Theorem 3.3 and Theorem 1.7). The proof includes a reproof of Teleman's Coulomb-gluing theorem and an appendix connecting the results to derived geometric Satake and relative Langlands duality.","tokens_in":26286,"tokens_out":7461,"duration_ms":69793,"significance":"If the main results hold, the paper introduces a genuinely new structural tool for Coulomb branches—expressing a Coulomb branch as a Hamiltonian reduction of a balanced product—and settles a conjecture of Bourget–Dancer–Grimminger–Hanany–Zhong. The consequences are substantial: finite generation of functions on T*(G/U_P) for GL_n and SL_n, symplectic singularities, holomorphic symplectic stratifications, and independence of the ordering of the partition. A notable strength is that the central functoriality theorem is proved in the paper, including a reproof of Teleman's theorem rather than merely importing it; the gluable hypothesis is stated explicitly, and the authors are honest in Remark 1.3 that they expect it to be unnecessary. The main weakness is that the key application in Theorem 3.3 does not verify the gluable hypothesis for the specific map used, leaving the proof of the headline conjecture conditional on an unstated check.","major_comments":[{"comment":"The isomorphism (3.5) is obtained by applying Theorem 1.1 to the map GL_Q_m/Gm → (GL_A_m × GL_A_n)/Gm, but the gluable condition of Definition 2.6 is never verified for this map. Lemma 3.1 verifies gluability for quiver dismemberments GL(V_Q) → GL(V_ˇQ), yet the map used here is a map to a product of two quiver gauge groups and is not shown to be a dismemberment covered by that lemma. Since Theorem 1.1 is conditional on gluability and Remark 1.3 concedes the hypothesis may be unnecessary, the proof of Theorem 1.7 currently relies on an unproven special case. The check is finite and explicit—the weights are coordinate characters—so the authors should either verify Definition 2.6 for this map, or prove a sufficiently general version of Theorem 1.1 without the gluability assumption, or explain precisely why Lemma 3.1 applies.","section":"§3.3, proof of Theorem 3.3"},{"comment":"The proof of exactness of the sequence (2.8) is compressed at a load-bearing point. The claim that conditions (1) and (2) of Definition 2.6 'exactly guarantee that these have no common factors after restriction to O(˜t_H)' is stated for Euler classes built from all weights, but condition (1) allows proportionality by an arbitrary rational α, and the Euler classes are products of linear factors with multiplicities. The authors should spell out why the rational proportionality cannot create a common factor when the sign condition (2) holds, and why the associated-graded exactness implies exactness of (2.8) without additional flatness or finiteness conditions on the Schubert filtration.","section":"§2.2, proof of Theorem 2.8"},{"comment":"The colimit argument used to prove that the map ˜f_{ς,N} is an isomorphism is abbreviated. The sentence 'Quotienting by fM(G) and using the fact that colimits commute with colimits, the fact that ς′`ς′`ς′ is an isomorphism immediately follows...' skips the details of why the quotient of the pushout by the fM(G)-action is the desired balanced product, and why the relevant colimits are well-behaved in the category of affine schemes. This is a central step in establishing Theorem 1.1, so it needs a fuller justification or a reference to a standard colimit argument.","section":"§2.4, proof of Theorem 2.13"}],"minor_comments":[{"comment":"The quiver diagrams for Q_⃗m and A_⃗m in Section 1.3 and Section 3.3 are barely legible in the plain-text rendering; a properly typeset diagram or an explicit description of the vertex and edge sets would greatly aid the reader.","section":"Throughout"},{"comment":"The notation ⟨ξ_i|μ⟩ for the pairing of a weight and a cocharacter is used without defining the pairing; please add a sentence explaining it is the natural perfect pairing between the character lattice and cocharacter lattice.","section":"Definition 2.6"},{"comment":"The balanced product in (1.5) is written with g(ˇQ,n)//G(ˇQ,n), while the general construction in (1.1) uses the notation g//G for the invariant adjoint quotient; the notation should be harmonized.","section":"Corollary 1.5"},{"comment":"There is a typo: 'necessarily changes' should be 'necessary changes'.","section":"Remark 2.11"},{"comment":"The proof of Lemma 3.5 relies on several imported results ([13, Theorem 2.11], [6], [26, Proposition 3.3]) and would benefit from a sentence explaining how the central isogeny pullback is compatible with the algebra-object base change, since this is the step that produces the isomorphisms in (A.7) for the GL_n and L cases.","section":"Appendix A.2"}],"recommendation":"major_revision","confidential_remarks":"The missing gluability check in §3.3 is the main obstacle; it appears to be a finite, explicit verification or a modest generalization of Lemma 3.1, so the paper is likely to be salvageable without changing its central claims. The self-citation [25] is published in IMRN and is not a circularity concern for the functoriality theorem itself. I would encourage the editor to request the authors to fill the gluability gap and expand the compressed proofs of Theorems 2.8 and 2.13 before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here is the short version: this is a serious paper with a genuinely new theorem, and the main application is attractive, but the proof as written has a gap at the last step. The functoriality theorem (Theorem 1.1) genuinely extends Teleman's gluing property to arbitrary maps H → G under a 'gluable' condition, and the authors reprove Teleman's theorem, so the central mechanism is self-contained. The quiver dismemberment result (Corollary 1.5) is new and gives a compositional calculus for Coulomb branches. The proof of the Bourget–Dancer–Grimminger–Hanany–Zhong conjecture for GL_n and SL_n is a substantial application, with a nice corollary on finite generation. The paper also includes a useful appendix connecting to geometric Satake and relative Langlands duality. Now the soft spots. The proof of Theorem 2.13 is compressed: the colimit argument is more of a sketch than a proof, and the associated-graded exactness in Theorem 2.8 is asserted rather than fully shown. A referee should ask for details, but these look repairable. More importantly, the proof of Theorem 3.3 applies Theorem 1.1 to the map GL_Qm/Gm → (GL_Am × GL_An)/Gm without verifying that this map is gluable. Lemma 3.1 covers quiver dismemberments, but this map goes to a product, not a dismemberment as covered by the lemma. The weights are coordinate characters, so the check is finite and explicit; but as written, the proof of isomorphism (3.5), and therefore of Theorem 1.7, is conditional on an unproven hypothesis. The final identification also imports the first author's prior result [25] for the last step. That is legitimate, since [25] is published, but it is load-bearing. The self-citation pattern is not a problem here; the gluing argument is otherwise independent. No circularity. Bottom line: this deserves peer review. The ideas are good, the gaps are explicit and probably fixable, and the payoff is high. If I were the editor I would send it out, with a referee instruction to demand the gluability check for the product map and a fuller proof of Theorem 2.13.","headline":"A real new functoriality theorem for Coulomb branches with a high-value application, but the proof of the headline conjecture is missing a gluability check that is likely an easy fix.","tokens_in":26857,"tokens_out":3157,"would_cite":true,"duration_ms":27791,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14L30","14M17","14B05"],"pacs":[],"model":"deepseek-v4-flash","headline":"For gluable maps H to G, the Coulomb branch of the restricted representation is the balanced product of the Coulomb branches for H and G; this proves the parabolic base affine space for GL_n or SL_n is a quiver Coulomb branch.","keywords":["Coulomb branches","quiver gauge theories","parabolic base affine space","partial implosions","Hamiltonian reduction","affine Grassmannian","symplectic singularities","transverse slices"],"falsifier":"Compute the balanced product and the Coulomb branch for the inclusion $\\mathrm{SL}_n \\subset \\mathrm{GL}_n$ with $N$ the direct sum of the standard representation and its dual: the weights $e_i$ and $-e_i$ restrict proportionally to the torus of $\\mathrm{SL}_n$ and have opposite signs on a cocharacter of $\\mathrm{GL}_n$, so the map is not gluable. If the two varieties are not isomorphic in this example, the gluable condition is necessary; if they are isomorphic, that supports the conjecture that the condition can be dropped.","tokens_in":25810,"feed_emoji":"🔗","tokens_out":17530,"duration_ms":142843,"temperature":0.7,"pith_summary":"The paper establishes a functoriality principle for Coulomb branches, the symplectic varieties that the standard Coulomb branch construction attaches to a reductive group and a representation. For a map $H \\to G$ of reductive groups and a representation of $G$ satisfying a mild 'gluable' condition, the Coulomb branch for the restricted representation of $H$ is canonically a balanced product of the Coulomb branch for $H$ with zero representation and the Coulomb branch for $G$, taken over the group scheme $M_C(G,0)$. This makes Coulomb branches behave as a kind of contravariant construction under Hamiltonian reduction. Two consequences follow: the affine closure of the cotangent bundle of the parabolic base affine space of $\\mathrm{GL}_n$ or $\\mathrm{SL}_n$ is a Coulomb branch of a quiver gauge theory, settling a standing conjecture, and the coordinate rings of these spaces are finitely generated.","feed_headline":"Parabolic base affine spaces proven to be quiver Coulomb branches","feed_subtitle":"A gluing formula for Coulomb branches settles the conjecture and proves finite generation of the coordinate rings.","key_machinery":"The load-bearing mechanism is the gluable condition on a map of pairs $(\\tilde{H},H) \\to (\\tilde{G},G)$ (Definition 2.6): no two weights $\\xi_1$, $\\xi_2$ of the representation $N$ restrict to proportional characters on the maximal torus $T_H$ and take opposite signs on some cocharacter of $G$. This condition ensures that the Euler classes $\\varphi_\\ell$ and $\\varphi_r$ attached to the two tautological bundles on the affine Grassmannian are coprime after restriction to the torus of $H$; coprimality makes the associated graded exact sequence (2.8) exact, and that is what forces the map from the balanced product to be an isomorphism. The argument runs through the Coulomb gluing property (Definition 2.3), a pushout description of the deformed Coulomb branch: its ring of functions consists of the functions on the $N=0$ Coulomb branch whose pullback by a rational automorphism $\\varpi_N$ stays regular, a property the paper proves by refining a known theorem for the triple $(G \\times G^{\\mathrm{sc}}_m, G, N)$.","core_discovery":"The central discovery is Theorem 1.1: if $\\varsigma: H \\to G$ is a map of complex reductive groups that is gluable for a representation $N$ of $G$, then there is a canonical $M_C(H,0)$-equivariant isomorphism $$M_C(H,0) \\$times^{{M_C(G,0)}}$ M_C(G,N) \\cong M_C(H,N)$$ of affine varieties over $\\mathfrak{h}//H$. The balanced product realizes the right side as the Hamiltonian reduction of the product of the $H$-Coulomb branch for the zero representation and the $G$-Coulomb branch for $N$. The proof upgrades a gluing description of Coulomb branch rings from the literature: the gluable condition says that the Euler classes of two tautological bundles on the affine Grassmannian have no common factor after restriction to the torus of $H$, which makes the associated graded of the relevant exact sequence exact. From this, the paper derives that the Coulomb branch of any loopless quiver is determined by Coulomb branches of two-vertex quivers, and that the affine closures $T^*(\\mathrm{GL}_n/U_P)$ and $T^*(\\mathrm{SL}_n/U_P)$ are quiver Coulomb branches, confirming the partial implosion conjecture.","pith_inferences":["If the gluable condition is shown unnecessary, the balanced product formula would apply to arbitrary maps of reductive groups and arbitrary quiver dismemberments, making the Hamiltonian-reduction description fully general; a natural test case is the $\\mathrm{SL}_n \\subset \\mathrm{GL}_n$ inclusion with $N = V \\oplus V^*$, where the condition fails.","The balanced product formula suggests a compositional structure on Coulomb branches: it behaves like a lax functor from reductive pairs to symplectic varieties, so the fission and dismemberment operations could be seen as instances of a single gluing law rather than separate constructions.","The gluing law connects to the conjectural S-duality framework for Hamiltonian spaces: if the S-dual of a Hamiltonian $G$-space is recovered from the Coulomb branch via the universal centralizer action, then the balanced product would give a geometric recipe for S-duals of cotangent-bundle spaces.","The proof shows that the gluable condition can be read off from weight combinatorics alone, so for any concrete quiver and dimension vector one can algorithmically decide which dismemberments are admissible, which would make the reduction to two-vertex quivers machine-checkable."],"forward_implications":["If a quiver $Q$ has no loops and every pair of parallel edges lifts to parallel edges in a dismemberment $\\hat{Q}$, the Coulomb branch $M(Q,n)$ is the balanced product of $M(G(Q,n))$ and $M(\\hat{Q},\\hat{n})$ over $M(G(\\hat{Q},\\hat{n}))$; iterating, every loopless quiver Coulomb branch is built from two-vertex quivers.","The affine closures $T^*(\\mathrm{GL}_n/U_P)$ and $T^*(\\mathrm{SL}_n/U_P)$ are quiver Coulomb branches (Theorems 1.7 and 3.3), so they have symplectic singularities, Gorenstein rational singularities, and a finite stratification into holomorphic symplectic subvarieties.","The coordinate rings of these affine closures are finitely generated, a nontrivial fact because each is the invariant ring of a finitely generated algebra under a unipotent group.","The affine closure $T^*(G/U_P)$ is independent of the ordering of the ordered partition $\\vec{m}$ of $n$.","Fission of a loopless quiver: the Coulomb branch of the fissioned quiver is the balanced product of the Coulomb branch of the original quiver with the Coulomb branch of the fissioned group over the original gauge group (Corollary 3.2)."],"supporting_citations":[{"why":"Lays out the Coulomb branch construction, its flatness, and the finite-generation result used in Corollary 1.8.","marker":"[14]"},{"why":"Supplies the algebra object whose cohomology is the Coulomb branch, used in the arguments for Lemma 3.5 and the appendix.","marker":"[13]"},{"why":"Gives the Coulomb gluing property for the triple $(G \\times G_m^{\\mathrm{sc}}, G, N)$, which the paper refines into the functorial upgrade.","marker":"[42]"},{"why":"Gives the action of $M_C(G,0)$ on $M_C(G,N)$ as schemes over $\\mathfrak{g}//G$, reproved and extended in Corollary 2.12.","marker":"[41]"},{"why":"States the partial implosion conjecture that Theorem 1.7 confirms for $\\mathrm{GL}_n$ and $\\mathrm{SL}_n$.","marker":"[10]"},{"why":"Computes the quotient $(K_{S(L)} \\times K_{\\mathrm{SL}_n})/J_{\\mathrm{SL}_n}$ as $T^*(\\mathrm{SL}_n/U_P)$, the final step in Theorem 3.3.","marker":"[25]"},{"why":"Sets conventions for quiver gauge theories and identifies many Coulomb branches with slices in the affine Grassmannian, used in the dismemberment and parabolic arguments.","marker":"[12]"},{"why":"Describes Coulomb branches of affine type A quivers as bow varieties, used in Lemma 3.7 to give an explicit model for the Coulomb branch of a leg.","marker":"[39]"}],"fun_headline_variants":["Coulomb branches glued, partial implosion conjecture proven","Quiver Coulomb branches built from two-vertex pieces","Gluing formula yields Coulomb branches for parabolic spaces","Finite generation proven via Coulomb branch gluing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing hypothesis is that no two weights of the representation restrict to proportional functions on the torus of $H$ and take opposite signs on some cocharacter of $G$; the authors themselves believe this 'gluable' condition is probably unnecessary, but the proof as written needs it to make the key exact sequence exact.","fun_headline_variants_meta":{"raw":{"variants":["Coulomb branches glued, partial implosion conjecture proven","Quiver Coulomb branches built from two-vertex pieces","Gluing formula yields Coulomb branches for parabolic spaces","Finite generation proven via Coulomb branch gluing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000188,"raw_usage":{"total_tokens":1348,"prompt_tokens":975,"completion_tokens":373,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":591,"completion_tokens_details":{"reasoning_tokens":311}},"tokens_in":591,"tokens_out":373,"duration_ms":4223,"temperature":1.0,"reasoning_tokens":311,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:30:31.949813+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the balanced product and the Coulomb branch for the inclusion $\\mathrm{SL}_n \\subset \\mathrm{GL}_n$ with $N$ the direct sum of the standard representation and its dual: the weights $e_i$ and $-e_i$ restrict proportionally to the torus of $\\mathrm{SL}_n$ and have opposite signs on a cocharacter of $\\mathrm{GL}_n$, so the map is not gluable. If the two varieties are not isomorphic in this example, the gluable condition is necessary; if they are isomorphic, that supports the conjecture that the condition can be dropped.","supporting_citations":[{"cited_title":"Partial implosions and quivers","cited_arxiv_id":null,"evidence_quote":"States the partial implosion conjecture that Theorem 1.7 confirms for $\\mathrm{GL}_n$ and $\\mathrm{SL}_n$."},{"cited_title":"The Cotangent Bundle of G/UP and Kostant–Whittaker Descent","cited_arxiv_id":null,"evidence_quote":"Computes the quotient $(K_{S(L)} \\times K_{\\mathrm{SL}_n})/J_{\\mathrm{SL}_n}$ as $T^*(\\mathrm{SL}_n/U_P)$, the final step in Theorem 3.3."},{"cited_title":"Cherkis bow varieties and Coulomb branches of quiver gauge theories of affine type A","cited_arxiv_id":null,"evidence_quote":"Describes Coulomb branches of affine type A quivers as bow varieties, used in Lemma 3.7 to give an explicit model for the Coulomb branch of a leg."}],"review_version":1}