Mathematics.

geometric methods

Geometric Representation Theory

Representation Theory90 minDifficulty10 out of 10

Overview

Geometric representation theory uses tools from algebraic geometry -- sheaves, D-modules, perverse sheaves, and derived categories -- to study representations of Lie algebras, algebraic groups, and related structures. The landmark Beilinson-Bernstein localization theorem (1981) establishes an equivalence between representations of a semisimple Lie algebra (the BGG category O) and D-modules on the flag variety. This geometric perspective revealed deep connections between representation theory, the geometry of Schubert varieties, and the topology of Lie groups.

Intuition

The flag variety G/B (quotient of a semisimple algebraic group by its Borel subgroup) is a smooth projective variety encoding the combinatorics of the Weyl group. Each weight lambda gives a line bundle L(lambda) on G/B; the Borel-Weil theorem says the global sections H^0(G/B, L(lambda)) form the irreducible representation L(lambda) when lambda is dominant. Going further, Beilinson-Bernstein shows ALL representations in the principal block of category O can be encoded as D-modules (differential operators) on G/B. Schubert varieties -- the orbit closures of the Borel -- have intersection cohomology that computes the Kazhdan-Lusztig polynomials.

Formal Definition

Definition

The flag variety is X = G/B for G a connected semisimple algebraic group and B a Borel subgroup. For a weight lambda in h*, define a line bundle L(lambda) on X via the character e^lambda of B. The Beilinson-Bernstein theorem: the global sections functor Gamma: D_lambda-mod -> U(g)_chi(lambda)-mod is an equivalence of categories when lambda is a regular dominant weight, where D_lambda is the sheaf of twisted differential operators on X with twist lambda.

X=G/B (flag variety)X = G/B \text{ (flag variety)}
Flag variety
H0(G/B,L(λ))L(λ) for dominant integral λH^0(G/B,\, \mathcal{L}(\lambda)) \cong L(\lambda) \text{ for dominant integral } \lambda
Borel-Weil theorem
Γ:Dλ-modU(g)χ(λ)-mod\Gamma: \mathcal{D}_\lambda\text{-mod} \xrightarrow{\sim} U(\mathfrak{g})_{\chi(\lambda)}\text{-mod}
Beilinson-Bernstein localization (dominant regular lambda)
[M(yλ):L(wλ)]=Py,w(1)[M(y\cdot\lambda):L(w\cdot\lambda)] = P_{y,w}(1)
KL polynomials from IC sheaves of Schubert varieties

Notation

NotationMeaning
G/BG/BFlag variety (complete flag manifold)
L(λ)\mathcal{L}(\lambda)Line bundle on G/B associated to weight lambda
Dλ\mathcal{D}_\lambdaSheaf of twisted differential operators
IC(Xw)IC(X_w)Intersection cohomology sheaf of Schubert variety X_w

Theorems

Theorem 1: Borel-Weil-Bott Theorem
ForGaconnectedreductivealgebraicgroupandlambdaanintegralweight:iflambdaisdominant,H0(G/B,L(lambda))istheirreduciblerepresentationL(lambda)andHi(G/B,L(lambda))=0fori>0.IfwlambdaisdominantforsomewintheWeylgroupoflengthl(w),thenHl(w)(G/B,L(lambda))=L(wlambda)andallothercohomologyvanishes.For G a connected reductive algebraic group and lambda an integral weight: if lambda is dominant, H^0(G/B, L(lambda)) is the irreducible representation L(lambda) and H^i(G/B, L(lambda)) = 0 for i > 0. If w*lambda is dominant for some w in the Weyl group of length l(w), then H^{l(w)}(G/B, L(lambda)) = L(w*lambda) and all other cohomology vanishes.
Theorem 2: Beilinson-Bernstein Localization
LetgbeasemisimpleLiealgebraoverC,X=G/Btheflagvariety,andlambdaadominantregularweight.ThelocalizationfunctorDelta:U(g)chi(lambda)mod>Dlambda(X)modandtheglobalsectionsfunctorGammaarequasiinverseequivalencesofabeliancategories.Underthisequivalence,Vermamodulescorrespondtodeltasheaves(directimagesfromSchubertcells),andirreduciblescorrespondtoICsheaves.Let g be a semisimple Lie algebra over C, X = G/B the flag variety, and lambda a dominant regular weight. The localization functor Delta: U(g)_{chi(lambda)}-mod -> D_lambda(X)-mod and the global sections functor Gamma are quasi-inverse equivalences of abelian categories. Under this equivalence, Verma modules correspond to delta-sheaves (direct images from Schubert cells), and irreducibles correspond to IC sheaves.
Theorem 3: Kazhdan-Lusztig via Intersection Cohomology
TheKazhdanLusztigpolynomialPy,w(q)equalsthePoincarepolynomialofthestalkoftheintersectioncohomologysheafIC(Xw)atapointintheSchubertcellCy:Py,w(q)=sumidimH2i(IC(Xw)y)qi.Inparticular,Py,w(1)=[M(ylambda):L(wlambda)]isthecompositionmultiplicityofVermamodules.The Kazhdan-Lusztig polynomial P_{y,w}(q) equals the Poincare polynomial of the stalk of the intersection cohomology sheaf IC(X_w) at a point in the Schubert cell C_y: P_{y,w}(q) = sum_i dim H^{2i}(IC(X_w)_y) * q^i. In particular, P_{y,w}(1) = [M(y*lambda):L(w*lambda)] is the composition multiplicity of Verma modules.

Worked Examples

  1. 1

    SL(2) acts on C^2. The Borel subgroup B = upper triangular matrices. The flag variety G/B = P^1 = CP^1 (the Riemann sphere), parametrizing lines in C^2.

    G/B=P1G/B = \mathbb{P}^1
  2. 2

    The weight n corresponds to the line bundle O(n) on P^1 (the n-th power of the tautological bundle).

  3. 3

    Borel-Weil: H^0(P^1, O(n)) = Sym^n(C^2)^* = the (n+1)-dimensional irreducible representation L(n) for n >= 0.

    H0(P1,O(n))L(n),n0H^0(\mathbb{P}^1, \mathcal{O}(n)) \cong L(n),\quad n \ge 0
  4. 4

    For n = -2: H^1(P^1, O(-2)) = C by Serre duality = L(-2) = trivial? Actually H^1(P^1, O(n)) = L(-n-2) for n <= -2. For n=1: standard rep L(1) = C^2 is H^0(P^1, O(1)).

✓ Answer

For SL(2), the flag variety is P^1, and global sections of O(n) give the (n+1)-dimensional irreducible L(n). This is the simplest instance of the Borel-Weil theorem.

Practice Problems

Hardfree response

Explain in broad strokes why the Beilinson-Bernstein theorem is significant: what does it gain over purely algebraic methods?

Common Mistakes

Common Mistake

Thinking geometric representation theory replaces algebraic methods entirely.

Geometric methods are powerful for certain classes of questions (KL polynomials, character formulas, structure of category O) but algebraic methods (Verma modules, universal enveloping algebras, PBW theorem) remain essential for concrete computations and for working in characteristic p where the geometry is more subtle.

Quiz

The Borel-Weil theorem states that for a dominant integral weight lambda:

Historical Background

The geometric approach began with Borel-Weil-Bott's construction (1954) of irreducible representations as cohomology of line bundles on flag varieties. Beilinson and Bernstein's 1981 localization theorem transformed the field by realizing representations as D-modules. Their proof of the Kazhdan-Lusztig conjecture via intersection cohomology of Schubert varieties was a landmark. Lusztig and Nakajima later developed geometric constructions of quantum group representations using quiver varieties.

  1. 1954

    Borel-Weil-Bott theorem: irreducibles as cohomology of line bundles on G/B

    Armand Borel, Andre Weil, Raoul Bott

  2. 1981

    Beilinson-Bernstein localization: g-modules = D-modules on flag variety

    Alexander Beilinson, Joseph Bernstein

  3. 1981

    Kazhdan-Lusztig conjecture proved via perverse sheaves

    Alexander Beilinson, Joseph Bernstein, Jean-Luc Brylinski, Masaki Kashiwara

  4. 1994

    Nakajima introduces quiver varieties to construct representations geometrically

    Hiraku Nakajima

Summary

  • Geometric rep theory uses algebraic geometry (sheaves, D-modules) to study representations of Lie algebras.
  • Borel-Weil: irreducibles L(lambda) = global sections of line bundles on the flag variety G/B.
  • Beilinson-Bernstein localization: g-mod (category O) = D-modules on G/B (for regular dominant lambda).
  • Kazhdan-Lusztig polynomials = Poincare polynomials of IC sheaves of Schubert varieties.

References

  1. BookHotta, R., Takeuchi, K., and Tanisaki, T. D-Modules, Perverse Sheaves, and Representation Theory. Birkhauser, 2008.
  2. BookHumphreys, J.E. Representations of Semisimple Lie Algebras in the BGG Category O. AMS, 2008.