🔢 mathematics

A relative Langlands dual realization of $T^*(G/K)$ and derived Satake

This paper establishes a relative Langlands dual realization of the cotangent bundle of a quasi-split symmetric space and extends the derived Satake equivalence to this setting by proving the equivalence for twisted affine Grassmannians and connecting these results to the geometric Langlands program on the twistor $\mathbb{P}^1$ .

Original authors: Tsao-Hsien Chen

Published 2026-01-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Tsao-Hsien Chen

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe of mathematics as a vast library filled with two very different types of books. On one shelf, you have Geometry, which deals with shapes, spaces, and how things move and twist (like the surface of a sphere or a twisted ribbon). On the other shelf, you have Representation Theory, which deals with abstract patterns, symmetries, and the "music" of groups (like how a kaleidoscope rearranges colors).

For a long time, mathematicians have been trying to find a secret dictionary that translates perfectly between these two languages. This is the heart of the Langlands Program, a grand theory that suggests every geometric shape has a matching musical pattern, and vice versa.

This paper, by Tsao-Hien Chen, is like discovering a new, previously missing chapter in that dictionary. It focuses on a specific, tricky type of shape called a Symmetric Space (think of it as a highly balanced, folded piece of paper) and its "shadow" or "dual" in the world of symmetries.

Here is a breakdown of what the paper achieves, using simple analogies:

1. The Main Discovery: A Perfect Mirror

The author looks at a specific geometric object: the cotangent bundle of a symmetric space.

The Analogy: Imagine a spinning top. The "symmetric space" is the shape of the top itself. The "cotangent bundle" is like a map showing every possible way that top could spin, wobble, or move at every single point on its surface. It's a complex, high-dimensional cloud of possibilities.
The Problem: Mathematicians wanted to know: "If we take this complex cloud of movement, what is its 'mirror image' in the world of symmetries?"
The Solution: Chen proves that for a specific class of these shapes (called "quasi-split"), the mirror image is exactly what you would expect if you took a "loop" version of the dual group.
The Metaphor: It's like taking a complex, twisted knot (the geometric shape) and realizing that its perfect mathematical twin is a specific type of braided rope made from a different material (the dual group). The paper shows that the "blueprint" for the knot's movement is identical to the "blueprint" for the rope's structure.

2. The "Loop" Trick

To make this translation work, the author uses a clever trick involving loops.

The Analogy: Imagine you have a piece of string. If you just look at the string, it's simple. But if you imagine the string as a loop that can be stretched, twisted, and moved through time (a "loop space"), it becomes incredibly complex.
The Innovation: The paper introduces a "twisted loop." Instead of just looping the string normally, you twist it as you go around.
The Result: By studying these twisted loops, the author finds a way to build a "dictionary" (called a Derived Satake Equivalence) that translates the complex geometry of the symmetric space directly into the language of sheaves (which are like layers of data or "sticky notes" attached to the loops).

3. The "Ring" of Information

The paper also talks about Ring Objects.

The Analogy: Think of a ring object as a special "key" or a "master recipe." In this mathematical world, you can take a geometric shape and turn it into a recipe. If you follow the recipe, you can reconstruct the shape's properties.
The Finding: The author shows that the "recipe" for the symmetric space's movement (the cotangent bundle) is built from the same ingredients as the "recipe" for the dual loop space. This confirms that the two worlds are not just similar, but fundamentally the same structure viewed from different angles.

4. Why This Matters (According to the Paper)

The paper doesn't talk about building bridges or curing diseases. Instead, it solves a deep puzzle in pure mathematics:

Bridging the Gap: It extends a famous rule (the Derived Satake Equivalence) that was previously only known for simple, "split" groups to a much more complicated, "quasi-split" world.
New Connections: It connects this geometry to a concept called the Twistor P1 (a specific type of geometric sphere used in theoretical physics and math) and suggests a new way to look at how these shapes relate to each other through a "Koszul duality" (a kind of mathematical inversion).
The "Bezrukavnikov" Link: It provides a new way to understand the "nilpotent cone" (a specific, singular shape where things break down or become unstable) by showing it matches up perfectly with a category of sheaves on the loop space.

Summary in One Sentence

This paper proves that the complex "map of all possible movements" for a specific type of folded geometric shape is mathematically identical to the "map of all possible twisted loops" for its dual partner, effectively writing a new, precise dictionary entry that allows mathematicians to translate seamlessly between these two distinct but related worlds of geometry and symmetry.

Technical Summary: A Relative Langlands Dual Realization of $T^*(G/K)$ and Derived Satake

Problem Statement
The paper addresses the problem of identifying the "relative dual" (or symplectic dual) variety $M_{\check{G}, \check{X}}$ associated with a spherical variety $\check{X}$ for a Langlands dual group $\check{G}$ . While the construction of such dual varieties via ring objects in the derived Satake category is well-established for split groups and specific cases (e.g., $\check{X} = \check{G}$ ), the paper investigates the inverse problem: given an affine Hamiltonian $G$ -variety $M$ of interest in representation theory, specifically the cotangent bundle $T^*(G/K)$ of a quasi-split symmetric space $X = G/K$ , can one identify a corresponding spherical variety $\check{X}$ for the dual group $\check{G}$ such that $M \cong M_{\check{G}, \check{X}}$ ?

The authors focus on the case where $G$ is a quasi-split symmetric space that is not necessarily split. In such cases, the naive candidate for the dual symmetric space (the quotient of the dual group by the dual of the symmetric subgroup) fails to produce the correct geometry. The paper seeks to resolve this by introducing "loop symmetric spaces" and establishing a derived categorical equivalence (Derived Satake) that generalizes known results for reductive groups to these quasi-split settings.

Methodology
The methodology combines geometric representation theory, the theory of loop groups, and the study of regular centralizer group schemes.

Geometric Setup and Kostant-Rallis Theory:
The authors utilize the Kostant-Rallis theory for symmetric spaces. They relate the cotangent bundle $T^*(G/K)$ to the regular centralizer group schemes $J_G$ and $J_H$ (where $H$ is a related split symmetric subgroup). Specifically, they establish an isomorphism between the coordinate ring of $T^*(G/K)$ and the ring of invariants of a specific quotient involving $G$ , $H$ , and the norm map between centralizer group schemes.
Twisted Loop Groups and Affine Grassmannians:
The core technical innovation involves the introduction of a "twisted" involution $\check{\theta}$ on the loop group $\check{G}(F)$ , defined by $\check{\theta}(\gamma)(t) = \check{\theta}_0(\gamma(-t))$ , where $\check{\theta}_0$ is a pinned involution of the dual group. This leads to the definition of the twisted loop group $\check{G}(F)^{\check{\theta}}$ and the twisted affine Grassmannian $\text{Gr}^{\check{\theta}} = \check{G}(F)^{\check{\theta}} / \check{G}(O)^{\check{\theta}}$ .
Derived Satake Equivalence for Twisted Grassmannians:
The authors prove a derived Satake equivalence for these twisted affine Grassmannians. They show that the derived category of $\check{G}(O)^{\check{\theta}}$ -equivariant sheaves on $\text{Gr}^{\check{\theta}}$ is equivalent to the category of perfect complexes on the shifted cotangent bundle of the split symmetric subgroup $H$ (specifically $T^*H[2]$ ). This extends the work of Zhu and others to the derived setting.
Ring Objects and Relative Duals:
Using the framework of BZSV (Ben-Zvi, Nadler, Sakellaridis) and BFN (Braverman, Finkelberg, Nakajima), the authors construct a ring object $A_{\check{G}, \check{\theta}}$ in the derived Satake category associated with the loop symmetric space $\check{X}^{\check{\theta}}(F) = \check{G}(F) / \check{G}(F)^{\check{\theta}}$ . They define the relative dual variety $M_{\check{G}, \check{\theta}}$ via the coordinate ring of this ring object.
Purity and Formality:
A crucial step involves proving formality and fully-faithfulness properties for $!$ -pure objects in the twisted derived Satake category. This allows the authors to compute the cohomology of these ring objects explicitly, linking them to the geometry of $T^*(G/K)$ .

Key Contributions and Results

Theorem 1.1 (Geometric Isomorphism): The authors establish a natural $G$ -equivariant isomorphism of rings:
$C[M_{\check{G}, \check{\theta}}] \cong C[T^*(G/K)]$
This identifies the relative dual of the loop symmetric space $\check{X}^{\check{\theta}}(F)$ with the cotangent bundle of the quasi-split symmetric space $G/K$ . This provides a concrete family of examples of relative Langlands duality for non-split groups.
Theorem 1.2 (Derived Satake for Loop Symmetric Spaces): The paper establishes a canonical equivalence between the derived Satake category for the loop symmetric space and the category of equivariant perfect complexes on the shifted cotangent bundle:
$D^b_{\check{G}(O)}(\check{X}^{\check{\theta}}(F))_e \cong \text{Perf}_G(T^*(G/K)[2])$
Here, the subscript $e$ denotes the subcategory generated by the dualizing sheaf of the base orbit under the Hecke action. This generalizes the derived Satake equivalence from reductive groups to quasi-split symmetric spaces.
Theorem 1.3 (Twisted Derived Satake and Purity):
1. A monoidal equivalence is proven for the twisted affine Grassmannian: $D^b_{\check{G}(O)^{\check{\theta}}}(\text{Gr}^{\check{\theta}}) \cong \text{Perf}_{H \times H}(T^*H[2])$ .
2. A fully-faithfulness result is established for $!$ -pure objects, providing a formula for the cohomology of tensor products involving the embedding of the twisted Grassmannian into the standard affine Grassmannian.
Bezrukavnikov Equivalence (Theorem 1.4): By combining the derived Satake equivalence with recent work on tamely ramified equivalence, the authors deduce a constructible realization of equivariant perfect complexes on the base change of the Grothendieck-Springer resolution along the moment map:
$D^b_{\check{I}}(\check{X}^{\check{\theta}}(F))_e \cong \text{Perf}_G((T^*(G/K) \times_{\mathfrak{g}^*} \tilde{\mathfrak{g}})[2])$
This serves as a version of the Bezrukavnikov equivalence for quasi-split symmetric spaces.
Extensions and Variants: The paper discusses variants of these results, including cases involving triality (for $\text{Spin}_8$ ), special parahoric subgroups (related to $SU_{2n+1}$ ), and subfields $F_n = \mathbb{C}((t^n))$ . These cases cover all twisted affine Grassmannians considered in prior literature (e.g., Zhu's work).

Significance and Claims
The paper claims to provide a "positive answer" to the question of whether the cotangent bundle of a quasi-split symmetric space arises as a relative dual variety. The significance lies in:

Generalization of Langlands Duality: It extends the relative Langlands duality framework, previously limited largely to split groups or specific spherical varieties, to the realm of non-split (quasi-split) symmetric spaces.
Categorical Realization: It provides a "constructible realization" of the category of perfect complexes on $T^*(G/K)$ in terms of sheaves on loop symmetric spaces, effectively categorifying the Plancherel formula for these spaces.
Connection to Geometric Langlands: The authors note that these results provide evidence for a "Twisted Relative Langlands Duality" conjecture, which posits a relationship between the derived category of the loop symmetric space and the category of constructible complexes on the moduli stack of real bundles on the twistor $\mathbb{P}^1$ .
Technical Foundation: The work establishes the derived Satake equivalence for twisted affine Grassmannians, a result of independent interest that relies on the study of nearby cycles functors and the formality of $!$ -pure objects.

The authors explicitly state that the results rely on $G$ being adjoint and $X$ being quasi-split. They also note that while the paper establishes the derived Satake equivalence, the quantum counterpart (relating to $(\mathfrak{g}, K)$ -modules) is reserved for forthcoming work. The paper does not claim to solve the full geometric Langlands program for these spaces but rather establishes a critical bridge between the geometry of symmetric spaces and the representation theory of loop groups via the lens of relative duality.