Rankin-Selberg duality via gluing
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "Rankin-Selberg duality via gluing".
Emmy: Here's the gist of the paper. The paper describes a gluing procedure used to describe the relative Langlands dual of hyperspherical Hamiltonian varieties, specifically focusing on the Rankin-Selberg case where m = n which recovers a result of Nakajima.
Ada: Description of the Relative Langlands Dual The relative Langlands dual of the hyperspherical (GLn × GLm)-variety T∗(Hom(Cm, Cn)) is described using a gluing procedure introduced by Ginzburg to describe the relative Langlands dual of T∗N, where N = Hom(Cm, Cn) is viewed as a spherical G-variety via the representation ρ given by (g, h) · a = gah−1.
Emmy: This dual is shown to be M(m,n), viewed as a (GLn × GLm)-variety via the action (1).
Ada: When m = n, this triangle part M is isomorphic to T∗(GLn×Cn) viewed as a hyperspherical Hamiltonian (GLn × GLn)-variety via the cotangent lift of the action on GLn × Cn by (g, h) · (a, v) = (gahT, gv).
Paper summary: Ada: So, we're looking at this paper "Rankin-Selberg duality via gluing" and it’s essentially about using a specific construction, one introduced by Ginzburg, to describe the relative Langlands dual of these hyperspherical Hamiltonian varieties. Emmy They take the general case for the variety T star Hom of C m to C n and show that this procedure results in an isomorphism to M m n, which they see as a triangle part of Cherkis-Nakajima-Takayama bow varieties. Ada The real point they make is that when m equals n, this construction actually recovers a known result from Nakajima and Takayama > Kolya That means the duality we are studying is equivalent to another well-known duality in the context of quiver gauge theories > Emmy It shows that these different mathematical tools can be used to describe these complex structures without needing to reinvent the wheel for every case > Kolya
Emmy: The authors also introduce an anti-symplectic involution on this Hamiltonian space, which they give an explicit formula for in the Rankin-Selberg case specifically > Ada This involution twists the action and connects it to a specific structure on T star GLn times Cn > Kolya It’s showing that the dual space has this particular symmetry under that transformation > Ada So in short, this paper establishes an isomorphism between two different geometric descriptions of these varieties, one involving gluing and the other involving the known result from Nakajima > Emmy
Kolya: I'm interested in how this connection back to Nakajima works. Is it just a structural match, or is there something deeper about the representation theory involved? Ada It seems like they’ve built a bridge between two different ways of looking at these dual spaces. The main result is that this gluing procedure gives you an isomorphism to M m n, which they identify with T star GLn times Cn when m equals n > Emmy
Emmy: And the real meat is what happens when they add that anti-symplectic involution. It twists the action in a very specific way, and that transformation connects it perfectly to the structure on T star GLn times Cn > Kolya What does that mean for understanding these types of Hamiltonian spaces? Does this technique help us get closer to understanding how these dualities work in general? Ada It gives us a concrete realization. They've shown that Spec A tau is isomorphic to T star GLn times Cn, and the action on it has this specific twisted symmetry > Emmy
Emmy: And they provide an explicit formula for that involution. That formula shows exactly how the coordinates change under that symmetry, which is pretty powerful for checking consistency > Kolya So, what's the big picture here? If we take this result—the isomorphism between the glued space and T star GLn times Cn—what does it suggest about these Langlands duals? Ada It suggests that even in these complicated setups, there are underlying geometric structures that are well-behaved and can be described precisely using these gluing tools > Emmy
Emmy: It validates the machinery of Ginzburg for describing relative Langlands duals > Kolya This whole result shows how to get a more concrete picture when m equals n, which is where the Rankin-Selberg case comes in > Ada
Ada: We've seen how they use this Ginzburg gluing method to describe the relative Langlands dual of these hyperspherical varieties, specifically focusing on that Rankin-Selberg case where m equals n. Emmy It boils down to an isomorphism between two geometric descriptions. They build this structure using a birational map and then show that it matches what Nakajima found for the triangle part of bow varieties > Kolya
Kolya: I'm interested in how this connects back to the original work they're referencing from Nakajima. Is it just a structural match, or is there something deeper about the representation theory involved? Ada It seems like they've built a bridge between two different ways of looking at these dual spaces. The main result is that this gluing procedure gives you an isomorphism to M m n, which they identify with T star GLn times Cn when m equals n > Emmy
Kolya: So, what's the big picture here? If we take this result—the isomorphism between the glued space and T star GLn times Cn—what does it suggest about these Langlands duals? Ada It suggests that even in these complicated setups, there are underlying geometric structures that are well-behaved and can be described precisely using these gluing tools > Emmy
Kolya: So, what's the big picture here? If we take this result—the isomorphism between the glued
Conclusion: Ada: So, we've just been looking at how Bruno da Silveira Dias used a Ginzburg gluing procedure to describe the relative Langlands dual of those hyperspherical varieties, especially focusing on that Rankin-Selberg case where m equals n.
Emmy: It really boiled down to showing an isomorphism between two different geometric descriptions. They took the construction from gluing and proved it matched what Nakajima found for those triangle parts of bow varieties.
Kolya: I'm still focused on that connection back to Nakajima’s original work, you know? Is it just a structural match, or is there something deeper about the representation theory driving this equivalence?
Ada: It looks like they built a bridge between two different ways of seeing those dual spaces. The main finding is that the gluing procedure gives you an isomorphism to M m n. When m equals n, they identify that with T star GLn times Cn.
Emmy: And what's really interesting is the anti-symplectic involution they introduce here for the Rankin-Selberg case. It twists the action in a very specific way, and it connects perfectly to the structure on T star GLn times Cn.
Kolya: So, what does that mean for understanding these Hamiltonian spaces in general? Does this technique help us get closer to how these dualities really work across the board?
Ada: It gives us a concrete realization. They showed that Spec A tau is isomorphic to T star GLn times Cn, and the action on it has this specific twisted symmetry. It's a very detailed map between the abstract gluing construction and that known geometric object.
Emmy: And they give an explicit formula for that involution. That formula shows exactly how the coordinates change under that symmetry, which is pretty powerful for checking consistency in these kinds of spaces.
Kolya: So, what's the big picture here then? If we take this isomorphism between the glued space and T star GLn times Cn, what does it suggest about these Langlands duals?
Ada: It suggests that even in these complicated setups, there are underlying geometric structures that are well-behaved and can be described precisely using these gluing tools. It validates the machinery of Ginzburg for describing relative Langlands duals.
Emmy: And the way they handled the involution, linking it to the structure on T star GLn times Cn, gives us a very specific lens through which to view these varieties.
Kolya: This whole result shows how we can get a more concrete picture when m equals n, which is where that Rankin-Selberg case really shines.
Ada: Exactly. It moves the discussion from abstract constructions to actual geometric objects we can analyze further. We've seen the core of the duality, now we're looking at its symmetry structure.
Emmy: And this work lays a foundation for using these gluing techniques in other related problems involving quiver gauge theories and representation theory.
math.RT
Submitted: 2026-07-25
Updated: 2026-11-05
Comments: 20 pages, v2: added section 5
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 78/100
The gist: The paper describes a gluing procedure used to describe the relative Langlands dual of hyperspherical Hamiltonian varieties…
Key concepts
- Relative Langlands Dual
- This is a mathematical object describing the symmetry properties of certain geometric spaces. The paper uses a gluing technique to construct this dual for hyperspherical varieties, which helps characterize the underlying representation theory in a specific context.
- Gluing Procedure ($ ho$ and $ au_ ho$)
- A Ginzburg-inspired method where two parts of a space are combined using a rational map and an automorphism. This procedure is used to define the coordinate ring of the dual variety and introduce an anti-symplectic involution that twists the group action by duality.
- $M(m,n)$
- This is a specific geometric variety that arises as the relative Langlands dual of certain structures. The paper shows that when $m=n$, this variety is isomorphic to the cotangent bundle of $GL_n imes C^n$ under a specific Hamiltonian action.
- Anti-symplectic Involution ($ au_ ho$)
- This is a transformation that twists the group action by duality. It is constructed using the gluing procedure and an anti-symplectic lift, leading to an isomorphism between the scheme Spec(A$ au$) and $T^*(GL_n imes C^n)$ with a modified action.
Terminology
Summary
The paper describes a gluing procedure used to describe the relative Langlands dual of hyperspherical Hamiltonian varieties, specifically focusing on the Rankin-Selberg case where m = n which recovers a result of Nakajima. p. 0
Description of the Relative Langlands Dual
The relative Langlands dual of the hyperspherical (GLn × GLm)-variety T∗(Hom(Cm, Cn)) is described using a gluing procedure introduced by Ginzburg to describe the relative Langlands dual of T∗N, where N = Hom(Cm, Cn) is viewed as a spherical G-variety via the representation ρ given by (g, h) · a = gah−1. This dual is shown to be M(m,n), viewed as a (GLn × GLm)-variety via the action (1). p. 1 When m = n, this triangle part M is isomorphic to T∗(GLn×Cn) viewed as a hyperspherical Hamiltonian (GLn × GLn)-variety via the cotangent lift of the action on GLn × Cn by (g, h) · (a, v) = (gahT, gv). p. 1
Construction of the Gluing and Automorphism
The gluing procedure involves defining a rational map γρ: gln × glm 99K GLn × GLm as given by (a, b) 7→ (χb(a),(−1)nχa(b)−1). This map is combined with the action on the Whittaker-twisted cotangent bundle WhGˇ to define a Gˇ-equivariant birational automorphism σρ: WhGˇ 99K WhGˇ defined as in (3). p. 2 The coordinate ring of the relative Langlands dual of T∗N is then described by Aσ:= O(WhGˇ) σ∗ ρ(f) ∈ O(WhGˇ). p. 3
Identification with Geometric Varieties
The main result proves that there is an isomorphism of Gˇ-schemes Spec(Aσ) = M(m,n). This is achieved by constructing two open imbeddings Φ1, Φ2: WhGˇ,→ M satisfying Φ2 = Φ1 ◦ σρ. p. 4 The proof relies on algebraic Hartogs’ lemma to show that the algebra homomorphism induced by the composition of these maps is surjective. p. 9
Incorporation of Duality Involution
The gluing procedure is modified so that the resulting Hamiltonian space is naturally equipped with an anti-symplectic involution. This involution, denoted as τρ:= θ ◦ σρ, twists the Gˇ-action by the duality involution ι. p. 1 In the Rankin-Selberg case m = n, this leads to an isomorphism Spec(Aτ) ∼= T∗(GLn × Cn) equipped with a specific (GLn × GLn)-action. p. 0
Explicit Formula for the Involution
The explicit formula for the involution τρ on the Hamiltonian space obtained by gluing in the Rankin-Selberg case is given by τρ(g, h, Cp, Cq) = (J(gT)−1q(CpTp−1S−1p,(−1)nJ(hT)−1p(CpTqS−1q, Cp, Cq). This involution is intertwined by the isomorphism with the action on T∗(GLn × Cn) given by (a, y, v, u) 7→ ((−1)n-1J(aT)−1J, JyT J, −JuT, −vT J). p. 3
Conclusion for the Rankin-Selberg Case
The final result establishes an isomorphism of affine schemes Spec(Aτ) ∼= T∗(GLn × Cn) = GLn × gln × Cn × (Cn). This isomorphism intertwines the Gˇ-action on Spec(Aτ) with the action (g, h) · (a, y, v, u) = (J(gT−1JaJh−1J, JhJyJh−1J, J(gT−1Jv, uJgT J). p. 3 Furthermore, the involution τρ on Spec(Aτ) is intertwined by this isomorphism with the involution on T∗(GLn × Cn) given by (a, y, v, u) 7→ ((−1)n-1J(aT−1J, JyT J, −JuT, −vT J). p. 0
Example for n = 1
For the case n = 1, the rational map γρ: C × C 99K C× × C× is given by (a, b) 7→ (a−b,(a−b)−1). The algebra Aσ can be computed directly as the subalgebra of O(WhGˇ) generated by the elements a, b, g, h−1, gh, h−1g−1,(a − b)g−1,(a − b)h. p. 8 Setting (gh, b, g,(a − b)g−1) =: (z, y, v, u) yields the isomorphism Spec(Aσ) ∼= C× × C × C × C. p. 4 The Gˇ-action on the right-hand side is (g, h) · (z, y, v, u) = (gzh, y, gv, uh−1). p. 9
Gluing via Involution
The map f 7→ θ∗(f) gives an isomorphism of rings Aτ → Aσ, and thus induces an isomorphism Spec(Aτ) → Spec(Aτ). Since θ(g · x) = ι(g) · θ(x), this isomorphism twists the Gˇ-action by the duality involution ι. p. 1
Example for GLn
In the case of GLn with the standard pinning, the duality involution is given explicitly by g 7→ J(gT−1J, where J is the antidiagonal matrix J = (0 0 · · · 0 1; 0 0 · · · 1 0;....;...;,.:.;,.:.). The anti-symplectic lift I to T∗GLn, identified via left-invariant trivialization with GLn × gln, is given by (g, a) 7→ (J(gT−1J, JaT J). p. 3 This descends to an anti-symplectic involution θ: WhGLn → WhGLn. p. 0 Under the identification WhGLn ∼= GLn × Sn, this involution is described as θ(g, Cp) = (J(gT−1S−1p, Cp). p. 3
Example for Universal Coulomb Branch
In the case of the universal Coulomb branch considered in [Gin25, Section 3.3], where N = Cn is the standard representation of GLn, the cotangent bundle T∗gln = gln × gln is realized as the gluing of two copies of T∗GLn. p. 3 The corresponding anti-symplectic involution of T∗gln is then (a, b) 7→ (−bT, aT). p. 0
Summary of Key Findings
The paper successfully uses the Ginzburg gluing procedure to describe the relative Langlands dual of T∗N and proves that this dual is isomorphic to M(m,n). The authors construct open imbeddings Φ1 and Φ2: WhGˇ,→ M whose images are M1 and M2, respectively, such that Φ2 = Φ1 ◦ σρ. p. 8 This construction leads to the isomorphism Spec(Aσ) ∼= M. p. 9 The final result shows that the resulting scheme Spec(Aτ) is isomorphic to T∗(GLn × Cn), equipped with a specific twisted Gˇ-action and an intertwining involution. p. 0
Pruning redundant reasoning tokens
The proof of Theorem 1.1 relies on the identification of the Coulomb branch of quiver gauge theories of affine type A with bow varieties, combined with [BFN19, Theorem 2. p. 111]. The proof is more straightforward than in Nakajima's work and uses a gluing procedure introduced by Ginzburg building on work of Teleman [Tel21] and Gannon-Webster [GW25].
How the training signal is built
The construction of the maps Φ1: WhGˇ → M1 and Φ2: WhGˇ → M2 involves defining linear maps Kp,q and Lp,q using the Bezoutian matrices Sp and Sq, which are related to the structure of Vp ⊗ Vp. These matrices satisfy specific matrix identities derived from the definition of Bezoutians. p. 0 Lemma 3.1 provides crucial matrix identities relating Kp,q and Lp,q. p. 6
Where the gains come from
The key technical steps involve showing that the stability conditions (15) and (16) hold by analyzing the properties of cyclic vectors for companion matrices. The proof that u = ξp(y2) involves induction on k to show g−1xyk2 = Ck p g−1x + X k j=1 Cj−1 p e1uyk−j2 = Ck p g−1x + X k j=1 ejuyk-j2. p. 0 This ultimately establishes the required stability conditions for the open imbeddings. p. 0 p. 0 p. 0 The Rankin-Selberg case m = n is handled by identifying M(n,n) with T∗(GLn × Cn) via a specific isomorphism. p. 0 p. 0 p. 0
The key technical steps involve showing that the stability conditions (15) and (16) hold by analyzing the properties of cyclic vectors for companion matrices. The proof that u = ξp(y2) involves induction on k to show g−1xyk2 = Ck p g−1x + X k j=1 Cj−1 p e1uyk-j2 = Ck p g−1x + X k j=1 ejuyk-j2. p. 0 This ultimately establishes the required stability conditions for the open imbeddings. The Rankin-Selberg case m = n is handled by identifying M(n,n) with T∗(GLn × Cn) via a specific isomorphism.
The key technical steps involve showing that the stability conditions (15) and (16) hold by analyzing the properties of cyclic vectors for companion matrices. The proof that u = ξp(y2) involves induction on k to show g−1xyk2 = Ck p g−1x + X k j=1 Cj-1 p e1uyk-j2 = Ck p g−1x + X k j=1 ejuyk-j2. p. 0 This ultimately establishes the required stability conditions for the open imbeddings. The Rankin-Selberg case m = n is handled by identifying M(n,n) with T∗(GLn × Cn) via a specific isomorphism.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved system can compute relative Langlands duals of hyperspherical Hamiltonian varieties by implementing Ginzburg's gluing procedure, allowing it to describe these duals as a
triangle part of Cherkis-Nakajima-Takayama bow varieties.
-
The system can identify the relative Langlands dual of the hyperspherical variety T∗N as M(m,n) and its isomorphism to T∗(GLn × Cn) in the Rankin-Selberg case, enabling it to recover a result of Nakajima.
-
The AI can naturally equip Hamiltonian varieties with an anti-symplectic involution by modifying the gluing procedure suggested by Ginzburg, providing an explicit formula for this involution in the Rankin-Selberg case.
-
The system can perform geometric constructions using algebraic Hartogs' lemma to prove that Spec(Aσ) is isomorphic to M(m,n), demonstrating how open imbeddings Φ1 and Φ2 relate via the birational automorphism σρ.
-
The improved system can verify the relationship between the maps Φ1 and Φ2 by checking specific matrix identities derived from Bezoutian matrices, such as
Kp,q = Sp
when m <= n.
Sources
- Relative Langlands Duality
- Functoriality of Coulomb branches
- Pointwise purity, derived Satake, and Symplectic duality
- S-dual of Hamiltonian $\mathbf G$ spaces and relative Langlands duality
Related papers
- Square root crystals and the square root of B(infinity)
- Finite Root Sums in Symmetrizable Kac-Moody Algebras
- Koszul duality for finite-dimensional Koszul algebras
- Standard modules and Dold--Kan correspondences for generalized Reedy categories
- What is Smoothness?
- Shifted quantum affine symmetric pairs: finite presentations and graded coideal structures