Proposition
For a character table,
- the rows as “weighted” orthogonal by Schur’s orthogonality relations;
- for columns and ,
- the matrix satisfies , where ;
- the complex conjugate of a row is a row of the character table by (possible the same);
- the complex conjugate of a column is also a column by (possible the same);
- (twisting) the pointwise product of a row starting with with any other row is again a row;
- the pointwise product of any two rows is a linear combination of rows of the character table with non-negative integer coefficients.
Example. Different groups with the same character table: and .
Frobenius-Schur Indicator
Definition
Frobenius-Schur indicator is defined as , where is an irreducible character.
Remark. For compact group and its irreducible character , define where is Haar measure with in Gordon Liebeck, real representation.
Theorem
Let be an irreducible character. Then
where is called of symmetric (rep. skew-symmetric) type if it has a non-zero symmetric (rep. skew-symmetric) -invariant form.
Definition
We say that is symmetric (rep. skew-symmetric) type if admits non-zero symmetric (rep. skew-symmetric) -invariant bilinear form. A bilinear map is called a bilinear form, and is called
- -invariant if ;
- symmetric if ;
- skew-symmetric if .
Remark. If has character , then is a character of . Then:
- if is irreducible, then is irreducible as well;
- is real-valued iff iff is self-dual.
Tensor Square
If is a vector space, then . If is basis of , then is basis of and
Define , then preserves and . Note that (rep. ) is the eigenspace of corresponding to eigenvalue (rep. ). It yields that and . Since , we can check that , and .
If , then
and so
Note that . Hence, if is self-dual, then and one of the following holds:
- and , which will correspond to symmetric type of ;
- and , which will correspond to skew-symmetric type of .
Lemma
Let be -module, then
- ;
- ;
- .
\begin{proof}
Since is finite, has finite order and so is diagonalizable for all . For a fixed and vector space , we have basis of eigenvectors of such that are corresponding eigenvalues. It follows that
Similarly, there is
and now we finish the proof.
\end{proof}
Now we are ready to prove ^df16b3.
\begin{proof}
Note that
by here.
\end{proof}
Exercise. Let be vector space, then is a vector space of bilinear forms on Show that . (I think it is easy, with .)
Now consider . Since , we have that is the space of symmetric bilinear forms on and is the space of skew-symmetric bilinear forms on . If is -module, then is also -module as well, by .
If is -invariant, then and so the space of -invariant bilinear form is . Recall that
and it yields that the space of -invariant bilinear form can be decomposed as the space of symmetric -invariant bilinear forms and the space of skew -invariant bilinear forms. In fact, for each , it can be written as where and . The decomposition above coincides with it and we also have names real, complex and quaternionic type for symmetric, complex and skew-symmetric.
Remark. From linear algebra, odd dimensional space cannot have a non-degenerate skew symmetric bilinear form. Hence, if is self-dual and of odd dimension, then is of symmetric type.
Example. Check for characters of and .
- Note that these two groups have the same character table.
- When , and and for .
- When , and and for , where is defined here.
Complexification
Note that -dimensional complex representation is just a complexification of -dimensional real representation.
If is a -module, then is a -module. That is, if is an -basis of , then is a -basis of and usually we write that with .
Observation. if is an irreducible -module, is not always irreducible. For example, is irreducible over but reducible over .
Proposition
If is simple, then it is of symmetric type.