Assignment III Solutions

Felipe Bedoya R
University Sergio Arboleda - Mathematics

October 30, 2019

1 [S]1.13.15.

Let us define the inner tensor product of two representations by (X ⊗Y ˆ )(g) = X(g) ⊗ Y (g)
Now we will show some properties. First of all, it is a representation of G. To begin with, it
is rather obvious that tensor product of matrices returns a matrix and that if the degree of X
ˆ )(e) = Idf ×df where e is the neutral element of G.
is d and the the degree of Y is f , then (X ⊗Y
These inner product provides invertible matrices since det(Xd×d ⊗ Yf ×f ) = det(X)d det(Y )f
and X, Y are also invertible.

Thus it is left for us to show is a group homomorphism. We want to show that

ˆ )(gh) = (X ⊗Y
(X ⊗Y ˆ )(g)(X ⊗Y
ˆ )(h) (1)

By properties of tensor product and since X and Y are representations, we have that

ˆ )(gh) = X(gh) ⊗ Y (gh) = X(g)X(h) ⊗ Y (g)Y (h) = (X(g) ⊗ Y (g))(X(h) ⊗ Y (h)) (2)
(X ⊗Y

And this is just (X ⊗Y ˆ )(g)(X ⊗Yˆ )(h). Now we will analyze the character of this repre-
sentation. Since it is given by tensor product of X(g) and Y (g) we have that the central
blocks of the resulting matrix are X(g)(i,i) Y (g). We also know that the trace of that compo-
P
nent is X(g)(i,i) T r(Y (g)), hence the character of the entire matrix is i X(g)(i,i) T r(Y (g)) =
T r(X(g))T r(Y (g)).Thus, the character of the inner tensor product is just the product of the
tow characters.

On the other hand, the following question naturally arises. If X and Y are two irre-
ducible representations, is the inner tensor also irreducible? The answer is no, consider the

1
irreducible representation X3 of S3 in the complete character table. Let X3 ⊗ X3 be the
inner product, and calculate hχ3 ⊗ χ3 , χ3 ⊗ χ3 i = 16 π∈S3 χ3 (π) ⊗ χ3 (π)χ3 (π −1 ) ⊗ χ3 (π −1 ) =
P

1 P 2 −1 2
6 π∈S3 (χ3 (π)) (χ3 (π )) = 3. Thus, this the inner tensor product is not irreducible.

However, if X(g) is a degree one representation, it readily follows that (X ⊗Y ˆ ) is

irreducible whenever Y is, since the resulting matrix is just a scalar multiple of Y , thus the
inner product is irreducible as well.

2 [S]1.13.18.

The main goal is to prove the induction is transitive, which is to say that for K ≤ H ≤ G
with G a finite group we get that X ↑G ∼ H G
K = (X ↑K ) ↑H . We will do it by showing equality
between characters.

Recall that if we induce the representation of any subgroup H of G, then the respective
character is inducted as follows:

Let φ be the representation of K, then φ ↑G 1

φ(x−1 gx).
P
K= |K| x∈G

Let us induce H with base on K and then G through H. We will note θ as the induced
character on H.

1 −1
We have that θ(g) = φ ↑H
P
K (g) = |K| h∈H φ(h gh). Then we induce the representa-
1 −1
tion on G with θ ↑G
P
H (g) = |H| x∈G θ(x gx). By combining these results we have that the
character of the induction on G is

1 X 1 X
φ(h−1 x−1 gxh) (3)
|H| x∈G |K| h∈H

Notice that for each h ∈ H the set of {x0 = xh} for all x ∈ G is equal to G itself. Since
that is true for each h ∈ H, then (4) becomes

1 1
φ((x0 )−1 g(x0 ))
X
|H| (4)
|H| |K| x0 ∈G

And this of course is equal to

2
 1 X
φ ↑G
K= φ(x−1 gx). (5)
|K| x∈G

Since the characters of the induction are equal for all g ∈ G we hence have proven they
are isomorphic.

3 [S]2.12.4.

Consider S(n−1,1) and the related tabloids. The following facts about this module and its
character hold:

• Following the notation of Sagan (2010), for each tableaux we get et = ~k − ~i where ~k
should be understood as the tabloid whose single-at-the-bottom element is k, and so
on for all the other elements. Then we have that the span of such polytabloids is

c1,1 (~1 − ~2) + . . . , +c1,n (~1 − ~n) + c2,1 (~2 − ~1) + c2,3 (~2 − ~3) + · · · + c2,n (~2 −~n) + · · · + cn,1 (~n −
~1) + · · · + cn,n−1 (~n − n − ~ 1)

Now we proceed to group terms so we multiply each ~k by a combination of ci,j , call it

c0k . Consider ci,j (~i − ~j), we know that ci,j will be a term of c0i and −ci,j will be a term
of c0j then when we perform c01 + c02 + · · · + c0n it will be zero since it contains ci,j and
−ci,j for each pair (i, j).

• This means dim(S(n−1,1) ) = n−1 thus we can choose a basis B = {(~2 − ~1), . . . (~n − ~1)}.
Now we extend this basis to {(~1 + ~2 + . . . +
~ n), (~2 − ~1), . . . (~n − ~1)} = {e~1 , e~2 , . . . e~n } and
we use it to calculate the character. Recall that with this basis we have that X(π)e~1 =
e~1 and by decomposition we have that χ(g) = 1 + χn−1,1 (g) thus χn−1,1 (g) = χ(g) − 1
and remember that χ(g) counts the number of fixed points of π and that it is invariant
under change of basis since it is equal between isomorphisms.

• Now we will show that S(n−1,1) is an irreducible representation by showing that

hχ(g), χ(g)i = 1.
We will base on an extension of the derangement to formula to count how many
(n−k)
permutations there are in Sn with k fixed points. The result is n! ( 1 + · · · + (−1)
k! 0! (n−k)!
).
Also recall that the inner product of χ and itself is equal to n! π∈Sn χ(g)χ(g −1 ) when
1 P

3
chi as we have proven before, counts the number of fixed points minus one. Therefore,
(n−k)
it all comes to show that n!1 nk=0 (k − 1)2 n! ( 1 + · · · + (−1)
P
k! 0! (n−k)!
) = 1 This holds whenever

n
1 X n! 1 (−1)(n−k)
k2 ( + · · · + )=2 (6)
n! k=0 k! 0! (n − k)!

n
1 X n! 1 (−1)(n−k)
k ( + ··· + )=1 (7)
n! k=0 k! 0! (n − k)!

n
1 X n! 1 (−1)(n−k)
1 ( + ··· + )=1 (8)
n! k=0 k! 0! (n − k)!

Those three can be proven through several ways, by using Burnside’s lemma, by using
Frobenius Reciprocity and by combinatorial arguments. We will prove them by induc-
tion on n and by Newton’s Binomial Theorem. We will show (6) but (7) and (8) can
be shown in a very similar manner.
Let n ≥ 2, the base case is simple and the equality to 1 can be readily seen. Now let
(n−k)
us proceed with the inductive step. Assume that n!1 nk=0 k 2 n! ( 1 + · · · + (−1)
P
k! 0! (n−k)!
)=
Pn k 1
1. Notice that the equation vanishes for k = 0 so we can set k=1 (k−1)! ( 0! + ··· +
(−1)(n−k)
(n−k)!
) = 1. We can add on both sides the adequate terms for us to have

n+1 n+1
X k 1 (−1)(n+1−k) X k (−1)(n+1−k)
( + ··· + )=1+ (9)
k=1 (k − 1)! 0! (n + 1 − k)! k=1 (k − 1)! (n + 1 − k)!

n−1
X 1 (−1)(n−1−k)
2+ (10)
k=0 k! (n − 1 − k)!

By both taking out the first term and rescaling the sum. By binomial theorem, the
last term of the expression vanishes.