Let $\rho:G\rightarrow\text{GL}_n(\mathbb{C})$ be a representation of a finite group and let $\chi_\rho$ be the corresponding character. If $\chi(e)>1$, then I want to show that $\chi_\rho\cdot\overline{\chi_\rho}$ is not an irreducible character. (Here, $\overline{\chi_\rho}$ denotes the pointwise conjugate of $\chi_\rho$ and $\cdot$ denotes the pointwise product.) I see that $\chi_\rho\cdot\overline{\chi_\rho}$ is self conjugate so it must be a real valued function, though I can't see how to use this.

A character is irreducible iff the inner product is $1$, so I want to show that $\langle \chi_\rho\cdot\overline{\chi_\rho},\chi_\rho\cdot\overline{\chi_\rho}\rangle\neq1$. Then I have the following: $$\langle \chi_\rho\cdot\overline{\chi_\rho},\chi_\rho\cdot\overline{\chi_\rho}\rangle=\frac{1}{|G|}\sum_{g\in G}\overline{\chi_\rho(g)\overline{\chi_\rho(g)}}\chi_\rho(g)\overline{\chi_\rho(g)}$$

$$=\frac{1}{|G|}\sum_{g\in G}\chi_\rho(g)\overline{\chi_\rho(g)}\chi_\rho(g)\overline{\chi_\rho(g)}=\frac{1}{|G|}\sum_{g\in G}|\chi_\rho(g)|^4$$ But I'm not sure how to continue from here. I tried considering the character as the sum of roots of unity: $$\chi_\rho(g)\overline{\chi_\rho(g)}=(w_1+\dots+w_n)(\frac{1}{w_1}+\dots+\frac{1}{w_n})$$ for $w_i$ an (order of $g$)th root of unity and $n=\chi_\rho(e)$. Then I can rewrite this as: $$n+\sum_{i<j}^n\frac{w_i}{w_j}+\frac{w_j}{w_i}=n+2\sum_{i<j}^n\text{Re}\left(\frac{w_i}{w_j}\right)$$ but I can't see any way to proceed.

  • 1
    $\begingroup$ $\chi_\rho\cdot\bar{\chi}_\rho$ is the character of $\rho \otimes \bar{\rho}$, which is the product of two representations of dimension>1 and so it is reducible. $\endgroup$ – fqq Apr 19 '16 at 15:31
  • $\begingroup$ Thank you. We haven't covered the tensor product of representations yet, so I think the problem should have a purely character theoretic solution. $\endgroup$ – Tachyon Apr 19 '16 at 15:55
  • 1
    $\begingroup$ Try to show that $[\chi\overline{\chi},1_G]=[\chi,\chi]>0$ for any character $\chi$. Since $\chi\overline{\chi}$ has dimension greater than $1$, you are done. $\endgroup$ – Levent Apr 19 '16 at 19:04
  • $\begingroup$ Thanks for your hint. I have attempted a solution below. $\endgroup$ – Tachyon Apr 19 '16 at 22:59

Following Levent's comment:

Let $\chi=\sum_i n_i\chi_i$ be an irreducible decomposition of $\chi$. Then $$(\chi,\chi)=\left(\sum n_i\chi_i,\sum n_j\chi_j\right)=\sum_{i,j}n_in_j(\chi_i,\chi_j)=\sum_{i,j}n_in_j\delta_{i,j}=\sum_i n_i^2>0$$

Then $(\chi\overline{\chi},1_g)>0$. If $(\chi\overline{\chi},1_g)>1$, then $\chi\overline{\chi}$ is reducible. If $(\chi\overline{\chi},1_g)=1$, then there must be some other irreducible character appearing in the decomposition since $\chi\overline{\chi}(e)=\chi(e)^2>1=1_g(e)$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.