Computing Global Characters

Transcription

1

2 Computing Global Characters

3 Computing Global Characters π: irreducible admissible representation of G

4 Computing Global Characters π: irreducible admissible representation of G Problem: Compute the (distribution) character θ π of π

5 Computing Global Characters π: irreducible admissible representation of G Problem: Compute the (distribution) character θ π of π Harish-Chandra: function on the regular semisimple set

6 Computing Global Characters π: irreducible admissible representation of G Problem: Compute the (distribution) character θ π of π Harish-Chandra: function on the regular semisimple set Roughly: fix H, θ π (g) = a(π,w)e wλ (g) (g)

7 Computing Global Characters π: irreducible admissible representation of G Problem: Compute the (distribution) character θ π of π Harish-Chandra: function on the regular semisimple set Roughly: fix H, θ π (g) = a(π,w)e wλ (g) (g) Problem: Compute a(π, w)

8 Why?

9 θ π determines π Why?

10 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic?

11 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic? Applications: the Langlands program, lifting, base change,...

12 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic? Applications: the Langlands program, lifting, base change,... Stability Application: Given a unipotent Arthur paramter Ψ, compute the Arthur packet Π ψ.

13 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic? Applications: the Langlands program, lifting, base change,... Stability Application: Given a unipotent Arthur paramter Ψ, compute the Arthur packet Π ψ. Need to compute: AV(π) (a set of real nilpotent orbits)

14 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic? Applications: the Langlands program, lifting, base change,... Stability Application: Given a unipotent Arthur paramter Ψ, compute the Arthur packet Π ψ. Need to compute: AV(π) (a set of real nilpotent orbits) (not just AV(Ann(π)) (a single complex nilpotent orbit))

15 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic? Applications: the Langlands program, lifting, base change,... Stability Application: Given a unipotent Arthur paramter Ψ, compute the Arthur packet Π ψ. Need to compute: AV(π) (a set of real nilpotent orbits) (not just AV(Ann(π)) (a single complex nilpotent orbit)) Not known...

16 Why? θ π determines π - e.g., how do you tell when two representations are isomorphic? Applications: the Langlands program, lifting, base change,... Stability Application: Given a unipotent Arthur paramter Ψ, compute the Arthur packet Π ψ. Need to compute: AV(π) (a set of real nilpotent orbits) (not just AV(Ann(π)) (a single complex nilpotent orbit)) Not known... use character theory to get some information (see

17 Theme: When can you encapsulate very complicated objects with surprisingly little data?

18 Theme: When can you encapsulate very complicated objects with surprisingly little data? Example: Reductive groups of rank m, and semisimple rank n:

19 Theme: When can you encapsulate very complicated objects with surprisingly little data? Example: Reductive groups of rank m, and semisimple rank n: pair of m n integral matrices (A,B)

20 Theme: When can you encapsulate very complicated objects with surprisingly little data? Example: Reductive groups of rank m, and semisimple rank n: pair of m n integral matrices (A,B) such that A B t is a Cartan matrix

21 Theme: When can you encapsulate very complicated objects with surprisingly little data? Example: Reductive groups of rank m, and semisimple rank n: pair of m n integral matrices (A,B) such that A B t is a Cartan matrix (A,B) (g t A,Bg 1 ) for g GL(m,Z)

22 Example: Here is complete information about representations of SL(2, R), including their characters. block: block 0(0,1): 0 [i1] 1 (2,*) 0 e 1(1,1): 0 [i1] 0 (2,*) 0 e 2(2,0): 1 [r1] 2 (0,1) 1 1 block: klbasis 0: 0: 1 1: 1: 1 2: 0: 1 1: 1 2: 1 5 nonzero polynomials, and 0 zero polynomials, at 5 Bruhat-comparable pairs.

23 History Inducted character formula, focus on the discrete series

24 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later)

25 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g)

26 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g) unique irreducible representation π = π(λ) satisfying: sgn(w)(wλ)( g) θ π (g) = D( g) Question: Formula for θ π on other Cartans?

27 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g) unique irreducible representation π = π(λ) satisfying: sgn(w)(wλ)( g) θ π (g) = D( g) Question: Formula for θ π on other Cartans? notoriously difficult

28 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g) unique irreducible representation π = π(λ) satisfying: sgn(w)(wλ)( g) θ π (g) = D( g) Question: Formula for θ π on other Cartans? notoriously difficult Herb:

29 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g) unique irreducible representation π = π(λ) satisfying: sgn(w)(wλ)( g) θ π (g) = D( g) Question: Formula for θ π on other Cartans? notoriously difficult Herb: (1) Stable sums of discrete series (two-structures)

30 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g) unique irreducible representation π = π(λ) satisfying: sgn(w)(wλ)( g) θ π (g) = D( g) Question: Formula for θ π on other Cartans? notoriously difficult Herb: (1) Stable sums of discrete series (two-structures) (2) Endoscopy

31 History Inducted character formula, focus on the discrete series Harish-Chandra: compact Cartan subgroup T, Λ character of T ρ (later) D( g) = (1 e α (g))e ρ ( g) unique irreducible representation π = π(λ) satisfying: sgn(w)(wλ)( g) θ π (g) = D( g) Question: Formula for θ π on other Cartans? notoriously difficult Herb: (1) Stable sums of discrete series (two-structures) (2) Endoscopy Other approaches (Schmid, Goresky-Kottwitz-MacPherson, Zuckerman,...)

32 Alternative Approach: All representations at once, using KLV polynomials

33 Alternative Approach: All representations at once, using KLV polynomials (atlas software)

34 Alternative Approach: All representations at once, using KLV polynomials (atlas software) Assume regular infinitesimal character λ

35 Alternative Approach: All representations at once, using KLV polynomials (atlas software) Assume regular infinitesimal character λ Theorem: Π(G,λ) = {(H,Λ) Λ Ĥ(R) ρ,dλ λ}/g(r)

36 Alternative Approach: All representations at once, using KLV polynomials (atlas software) Assume regular infinitesimal character λ Theorem: (H,Λ) Π(G,λ) = {(H,Λ) Λ Ĥ(R) ρ,dλ λ}/g(r) { I(H, Λ) standard (induced) module π(h, Λ) irreducible Langlands quotient

37 Fix H, +,

38 Fix H, +,ρ = α,h ρ

39 Fix H, +,ρ = α,h ρ D( +, g) = (1 e α (g))e ρ ( g) ( g H(R) ρ )

40 Fix H, +,ρ = α,h ρ D( +, g) = (1 e α (g))e ρ ( g) ( g H(R) ρ ) H(R) + = {g H e α (g) > 1 (α real)}

41 Fix H, +,ρ = α,h ρ D( +, g) = (1 e α (g))e ρ ( g) ( g H(R) ρ ) H(R) + = {g H e α (g) > 1 (α real)} θ π (h) = a(π, +,Λ)Λ( g) D( +, g) (g H(R) + )

42 Fix H, +,ρ = α,h ρ D( +, g) = (1 e α (g))e ρ ( g) ( g H(R) ρ ) H(R) + = {g H e α (g) > 1 (α real)} θ π (h) = a(π, +,Λ)Λ( g) D( +, g) (g H(R) + ) (drop + )

43 Harish-Chandra s character formula for discrete series + induced character formula :

44 Harish-Chandra s character formula for discrete series + induced character formula : Proposition: Formula for θ I(H,Λ) on H(R): W θ I(H,Λ) (h) = R sgn(w)(wλ)(h) (h H(R) + ) D(h) W R = W(G(R),H(R)) W(G,H)

45 Harish-Chandra s character formula for discrete series + induced character formula : Proposition: Formula for θ I(H,Λ) on H(R): W θ I(H,Λ) (h) = R sgn(w)(wλ)(h) (h H(R) + ) D(h) W R = W(G(R),H(R)) W(G,H) Sue me:

46 Harish-Chandra s character formula for discrete series + induced character formula : Proposition: Formula for θ I(H,Λ) on H(R): W θ I(H,Λ) (h) = R sgn(w)(wλ)(h) (h H(R) + ) D(h) W R = W(G(R),H(R)) W(G,H) Sue me: I m surpressing an irksome sign

47 Harish-Chandra s character formula for discrete series + induced character formula : Proposition: Formula for θ I(H,Λ) on H(R): W θ I(H,Λ) (h) = R sgn(w)(wλ)(h) (h H(R) + ) D(h) W R = W(G(R),H(R)) W(G,H) Sue me: I m surpressing an irksome sign Corollary: Γ Ĥ(R) ρ: a(i(h,λ),γ) = { ±1 Γ = wλ 0 otherwise

48 Question: Formula for θ I(H,Λ) on other Cartan subgroups?

49 Question: Formula for θ I(H,Λ) on other Cartan subgroups? Theory of leading terms (growth of matrix coefficients):

50 Question: Formula for θ I(H,Λ) on other Cartan subgroups? Theory of leading terms (growth of matrix coefficients): If (*) Re dλ,α 0 for all α + then Λ occurs in I(H,Λ) and the character formula for no other standard module:

51 Question: Formula for θ I(H,Λ) on other Cartan subgroups? Theory of leading terms (growth of matrix coefficients): If (*) Re dλ,α 0 for all α + then Λ occurs in I(H,Λ) and the character formula for no other standard module: Theorem: Fix (H, Λ) satisfying (*): { a(i(h,λ ±1 (H,Λ) (H,Λ ) ),Λ) = 0 otherwise

52 {π(h,λ)} and {I(H,Λ)} are both bases of the Grothendieck group

53 {π(h,λ)} and {I(H,Λ)} are both bases of the Grothendieck group I = m(i,π)i

54 {π(h,λ)} and {I(H,Λ)} are both bases of the Grothendieck group I = m(i,π)i(multiplicity formula)

55 {π(h,λ)} and {I(H,Λ)} are both bases of the Grothendieck group I = m(i,π)i(multiplicity formula) π = M(I,π)I (character formula)

56 {π(h,λ)} and {I(H,Λ)} are both bases of the Grothendieck group I = m(i,π)i(multiplicity formula) π = M(I,π)I (character formula) This is precisely what is computed by the Kazhdan-Lustig-Vogan polynomials (the klbasis command)

57 Corollary: Assuming (*), a(π,λ) = ±M(I(H,Λ),π)

58 Corollary: Assuming (*), a(π,λ) = ±M(I(H,Λ),π) = ±P I,π (1) (KLV polynomial)

59 Corollary: Assuming (*), a(π,λ) = ±M(I(H,Λ),π) = ±P I,π (1) (KLV polynomial) General Λ: use coherent continuation (wgraph command)

60 Corollary: Assuming (*), a(π,λ) = ±M(I(H,Λ),π) = ±P I,π (1) (KLV polynomial) General Λ: use coherent continuation (wgraph command) a(π,w Λ) = ±M(I(H,Λ),w 1 π)

61 Corollary: Assuming (*), a(π,λ) = ±M(I(H,Λ),π) = ±P I,π (1) (KLV polynomial) General Λ: use coherent continuation (wgraph command) a(π,w Λ) = ±M(I(H,Λ),w 1 π) Conclusion: KLV-polynomials explicit formulas for all a(π, Λ)

62 Example: Sp(4, R) 0( 0,6): 0 [i1,i1] 1 2 ( 4, *) ( 5, *) 0 e 1( 1,6): 0 [i1,i1] 0 3 ( 4, *) ( 6, *) 0 e 2( 2,6): 0 [ic,i1] 2 0 ( *, *) ( 5, *) 0 e 3( 3,6): 0 [ic,i1] 3 1 ( *, *) ( 6, *) 0 e 4( 4,5): 1 [r1,c+] 4 9 ( 0, 1) ( *, *) 1 1 5( 5,4): 1 [C+,r1] 7 5 ( *, *) ( 0, 2) 2 2 6( 6,4): 1 [C+,r1] 8 6 ( *, *) ( 1, 3) 2 2 7( 7,3): 2 [C-,i1] 5 8 ( *, *) (10, *) 2 1,2,1 8( 8,3): 2 [C-,i1] 6 7 ( *, *) (10, *) 2 1,2,1 9( 9,2): 2 [i2,c-] 9 4 (10,11) ( *, *) 1 2,1,2 10(10,0): 3 [r2,r1] ( 9, *) ( 7, 8) 3 2,1,2,1 11(10,1): 3 [r2,rn] ( 9, *) ( *, *) 3 2,1,2,1

63 0: 0: 1 9: 0: 1 1: 1 1: 1: 1 2: 1 3: 1 2: 2: 1 4: 1 5: 1 3: 3: 1 6: 1 9: 1 4: 0: 1 1: 1 10: 0: 1 4: 1 1: 1 2: 1 5: 0: 1 3: 1 2: 1 4: 1 5: 1 5: 1 6: 1 6: 1: 1 7: 1 3: 1 8: 1 6: 1 9: 1 10: 1 7: 0: 1 1: 1 11: 2: q 2: 1 3: q 4: 1 9: 1 5: 1 11: 1 7: 1 8: 0: 1 1: 1 3: 1 4: 1 6: 1 8: 1

64 R R : Irreducible Modules π (2,1) (1, 2) (2, 1) ( 1, 2) (1, 2) ( 2, 1) ( 1, 2) ( 2, 1) π(0) 1,1 1,0 0,1 π(1) 1,1 1,0 0,1 π(2) 1, 0 1, 0 π(3) 1, 0 1, 0 π(4) 1,1-1,-1-1,1 1,-1 π(5) 1,0-1,0-1,0 1,0 π(6) 1,0-1,0-1,0 1,0 π(7) 1,0-1,0-1,0 1,0 1,0-1,0 π(8) 1,0-1,0-1,0 1,0 1,0-1,0 π(9) 1,1-1,-1-1,1 2,0 1,-1-2,0 π(10) 1,0-1,0-1,0 1,0 1,0-1,0-1,0 1,0 π(11) 0,1 0,-1 0,1 1,0 0,-1-1,0-1,0 1,0