Let $V$ a $\mathbb KG$-module such that the character of $V$ is such that $\chi_V(g)=0$ for all $g\in G\setminus \{1\}$. Show that there is an $m$ s.t. $$V=\underbrace{\mathbb KG\oplus\cdots\oplus \mathbb KG}_{m\text{ times}}.$$

To me, $\chi_V$ is the regular representation. I know that $$\mathbb KG=U_1\oplus\cdots\oplus U_n$$ where $U_i$ are irreducible submodule of $\mathbb KG$. So, $$\underbrace{\mathbb KG\oplus\cdots\oplus \mathbb KG}_{m\text{ times}}=U_1^{\oplus m}\oplus \cdots \oplus U_n^{\oplus m},$$ but how can I prove that it's $V$ ?

  • $\begingroup$ Are you familiar with the usual inner product on the characters? $\endgroup$ – Tobias Kildetoft Jun 20 '16 at 19:41
  • $\begingroup$ @TobiasKildetoft That approach only works if $\mathbb{K}$ is of characteristic zero. The OP does not say anything about that (although I highly doubt that the statement will hold in positive characteristics). $\endgroup$ – Batominovski Jun 20 '16 at 21:28
  • $\begingroup$ @Batominovski Indeed, even making sense of it becomes less trivial (what should be meant by character for example). $\endgroup$ – Tobias Kildetoft Jun 21 '16 at 4:59

Edit As @Batominovski and @TobiasKildetoft remarked, one should assume characteristic zero here. In positive characteristic $p$ and non-trivial $G$, the statement is indeed not true, at least if one takes the naive definition of characters: For example, the $p$-fold sum of the trivial representation satisfies the assumption but is not of the form $KG^{\oplus m}$ since the latter is not trivial.

For any irreducible character $\chi_U$ you have ${\mathbb Z}\ni \langle\chi_V,\chi_U\rangle = \frac{\chi_V(1)\chi_U(1)}{|G|}=\frac{\dim(V)}{|G|}\dim(U)$. In particular, taking $U$ to be the trivial representation shows $\frac{\dim(V)}{|G|}=:m\in {\mathbb N}$, and so any irreducible representation $U$ of $G$ occurs exactly $m \dim(U)$ times in $V$. Since $KG=\bigoplus_U U^{\oplus\dim U}$, it follows that $V\cong KG^{\oplus m}$.

  • $\begingroup$ In positive characteristic, one would need to specify what is even meant by a character (just taking traces loses too much information to be something one usually does). $\endgroup$ – Tobias Kildetoft Jun 21 '16 at 7:41
  • $\begingroup$ @TobiasKildetoft Thanks, I made a comment. $\endgroup$ – Hanno Jun 21 '16 at 10:37

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.