Murphy elements <math> <mo>ℂ</mo><msub> <mi>A</mi><mi>k</mi> </msub><mfenced> <mi>r</mi><mi>p</mi><mi>n</mi> </mfenced> </math>

Murphy elements ℂ Ak rpn

Arun Ram
Department of Mathematics and Statistics
University of Melbourne
Parkville, VIC 3010 Australia
aram@unimelb.edu.au
and

Department of Mathematics
University of Wisconsin, Madison
Madison, WI 53706 USA
ram@math.wisc.edu

Last updates: 20 June 2010

Murphy elements ℂ Ak rpn

Let S⊆ 12…k and let I⊆S∪ S' . Define bS , dI ∈ Ak by bS = S∪ S' , l l' l∉S   and   d I⊆S = I, Ic l l' l∉S . For example, in A9 , if S= 2458 and I= 24 4' 58 then

bS =

and

dI =

Note that dI = d Ic , d S∪ S' = d∅ = bS , d l l' = d l l' c = b S- l . For k∈ 1 2 ℤ ≥0 and r∈ ℤ >0 , Z k,r = n 2 + ∑ S ≥1 -1 S n-1 bS + ∑ κ I =0 modr -1 κ I dI - bS , where the outer sum is over S⊆ 12…k such that S≠∅ and the inner sum is over I⊆S∪ S' such that dI ∈ Ak r1n and dI ≠ bS .

  1. For n,r∈ ℤ ≥0 , κ r,n = 1 r ∑ m=0 r-1 ∑ 1≤i<j≤n tim tj -m s ij is a central element of ℂ G r,p,n . κ r,n = ∑ b∈λ c b ,as operators on   G r,p,n λj , the irreducible G r,p,n -module indexed by λj , where λ= λ 0 … λ r-1 is a multipartition with n boxes.
  2. Let n,k∈ ℤ ≥0 . Then, κ r,n = Z k,r   and   κ n-1 = Z k+ 1 2 r ,as operators on   V ⊗k .
  3. Let n∈ℂ,k∈ 1 2 ℤ ≥0 . Then Zk r is a central element of ℂ Ak rpn , and, if n∈ℂ is such that ℂ Ak rpn is semisimple then Zk r = ∑ b∈λ c b ,as operators on   Akλ , where Akλ is the irreducible ℂ Ak rpn -module indexed by λ.

Proof.

  1. The element κn is the class sum corresponding to the conjugacy class of the element s12 in G r,p,n and thus κ r,n is a central element of ℂ G r,p,n . The constant by which κ r,n acts on G r,p,n λj follows from theorem???.
  2. The first statement follows from parts (a) and (b) and Theorems 3.6 and 3.22 as follows. By theorem 3.6, ℂ Ak n ≅ End Sn V ⊗k if n≥2k. Thus, by Theorem 3.22, if n≥2k then Zk acts on the irreducible ℂ Ak n -module Akλ n by the constant given in the statement. This means that Zk is a central element of ℂ Ak n for all n≥k. Thus, for n≥2k,d Zk = Zk d for all diagrams d∈ Ak . Since the coefficients in d Zk (in terms of the basis of diagrams) are polynomials in n, it follows that d Zk = Zk d for all n∈ℂ.

    If n∈ℂ is such that ℂ Ak n is semisimple let χ ℂ Ak n λ be the irreducible characters. Then Zk acts on Akλ n by the statement, therefore it is a polynomial in n, determined by its values for n≥2k.

    The proof of the second statement is completely analogous using ℂ A k+ 1 2 , S n-1 , and the second statement in part (b).

  3. Then 2 κn v i1 ⊗…⊗ v ik = 1 r ∑ m=0 r-1 ∑ i,j=1,i≠j n tim tj -m s ij v i1 ⊗…⊗ t im tj -m s ij v ik = 1 r ∑ m=0 r-1 ∑ i,j=1,i≠j n 1- E ii - E jj + tim E ij + tj -m E ji v i1 ⊗…⊗ 1- E ii - E jj + tim E ij + tj -m E ji v ik . Expanding this sum, let c S,I = ∏ l∈ Sc δ il i l' -1 # l l' ⊆I +# l l' ⊆ Ic ξ m l∈I l' ∈ Ic -# l∈I l∈ Ic ∏ l∈I δ il i ∏ l∈ Ic δ il j and c S,I = ∏ l∈ Sc δ il i l' -1 # l l' ⊆I +# l l' ⊆ Ic ξ m l∈I l' ∈ Ic -# l∈I l∈ Ic ∏ l∈I δ il i ∏ l∈ Ic δ il j so that 2 κn v i1 ⊗…⊗ v ik is equal to 1 r ∑ S⊆ 1…k ∑ i 1' ,…, i k' ∑ m=0 r-1 ∑ i,j=1,i≠j n ∑ I⊆S∪ S' c S,I v i 1' ⊗…⊗ v i k' = 1 r ∑ S⊆ 1…k ∑ i 1' ,…, i k' ∑ m=0 r-1 ∑ I⊆S∪ S' ∑ i,j=1 n c S,I v i 1' ⊗…⊗ v i k' - ∑ i,j=1 n c S,I ' v i 1' ⊗…⊗ v i k' . Here Sc ⊆ 1…k corresponds to the tensor products where 1 is acting, I⊆S∪ S' corresponds to the tensor position that must equal i, and Ic corresponds to the tensor positions that must equal j.

    When S =0 the set I is empty and the sum in (???) is equal to n2 -n v i1 ⊗…⊗ v ik since c S,I = c S,I ' = ∏ l∈ 1…k δ il i l' . Assume S ≥1 and separate the sum according to the cardinality of I. Note that the sum for I is equal to the sum for Ic since the whole sum is symmetric in i and j. When I=S∪ S' , c S,I = c S,I ' = ∏ l∈ Sc δ il i l' -1 S ∏ <S∪ S' δ il i and the sum in (???) is equal to 1 r ∑ m=0 r-1 ∑ i 1' ,…, i k' n-1 ∑ i=1 n c S,I v i 1' ⊗…⊗ v i k' = n-1 -1 S bS v i1 ⊗…⊗ v ik . We get a similar contribution from the sum of the terms with I=∅.

    For each of the remaining subsets I⊆S∪ S' the sum in (???) contributes 0 when κ I ≠0 modr and -1 # l l' ⊆I +# l l' ⊆ Ic d I⊆S v i1 ⊗…⊗ v ik = -1 S +κ I d I⊆S - bS v i1 ⊗…⊗ v ik . when κ I =0 modr .

    For the second statement, since 1- E ii - E jj + E ii E jj vn = 0, if i=n  or  j=n, vn , if otherwise, 2 κ n-1 v i1 ⊗…⊗ v ik ⊗ vn = 1 r ∑ m=0 r-1 ∑ i,j=1,i≠j n-1 tim tj -m s ij v i1 ⊗…⊗ v ik ⊗ vn = 1 r ∑ m=0 r-1 ∑ i,j=1,i≠j n tim tj -m s ij v i1 ⊗…⊗ tim tj -m s ij v ik ⊗ 1- E ii - E jj + E ii E jj vn = 1 r ∑ m=0 r-1 ∑ i≠j s ij v i1 ⊗…⊗ v ik ⊗ vn + 1 r ∑ m=0 r-1 ∑ i,j=1,i≠j n 1- E ii - E jj + ξm E ij + ξ -m E ji v i1 ⊗… …⊗ 1- E ii - E jj + ξm E ij + ξ -m E ji v ik ⊗ - E ii - E jj vn = 1 r ∑ m=0 r-1 ∑ i,j=1 n tim tj -m s ij v i1 ⊗…⊗ tim tj -m s ij v ik ⊗ E ii E jj vn . The first sum is known to equal 2 κn v i1 ⊗…⊗ v ik and is known by the computation proving the first statement, and the last is 0 since i≠j. The middle sum is treated exactly as in (???) except that now the sum is over S such that k+1∈S and I such that k+1 k+1 ' ⊆I or k+1 k+1 ' ⊆ Ic . □

∑ i=1 n tim v i1 ⊗…⊗ v ik = ∑ i=1 tim v i1 ⊗…⊗ tim v ik = ∑ i=1 1- E ii + ξm E ii v i1 ⊗…⊗( 1- E ii + ξm E ii = ∑ i1' ,…, ik' ∑ I⊆ 1…k ξm -1 I ∏ l∈ Ic δ il i l' ∏ l∈ Ic δ il i δ i l' i v i 1' ⊗…⊗ v i k' = ∑ I⊆ 1…k ξm -1 I bI = n+k ξm -1 + ∑ I ≥2 ξm -1 I bI = n-k+k ξm + ∑ I ≥2 ξm -1 I bI .

References [PLACEHOLDER]

[BG] A. Braverman and D. Gaitsgory, Crystals via the affine Grassmanian, Duke Math. J. 107 no. 3, (2001), 561-575; arXiv:math/9909077v2, MR1828302 (2002e:20083)

page history