Suppose I have a representation $V$ of a finite group $G$, and decompose $V$ into $V = V_1 \oplus V_2$ where $V_1$ and $V_2$ are irreducible representations.

If I understand correctly, it is known that for any other decomposition $V = W_1 \oplus W_2$ into irreducibles $W_1, W_2$, we must have $V_1 \cong W_1$ and $V_2 \cong W_2$ (after reordering if necessary). However, I am having trouble seeing why this must be true.

I believe the idea is to apply Schur's Lemma to the identity map $I \colon V \to V$. The image $I(V_1)$ must be isomorphic to $V_1$ by Schur's Lemma, but I don't see how we can be sure that this image can be identified with one of the $W_i$ as opposed to some new subrepresentation $U \subset W_1 \oplus W_2$.


Let $W$ be an irreducible subrepresentation of $V_1 \oplus V_2$, where $V_1$ and $V_2$ are not isomorphic. Consider the corresponding projection maps $W \to V_1, W \to V_2$. By Schur's lemma, these maps are either zero or isomorphisms. If $W \to V_1$ is an isomorphism, then $W \to V_2$ must be zero, so $W$ must in fact be the subrepresentation $V_1 \oplus 0$. Similarly for if $W \to V_2$ is an isomorphism.

  • $\begingroup$ Thank you, that was very clear. I still wonder however about the case when $V_1$ and $V_2$ are isomorphic. $\endgroup$ – GoatsRule Jan 17 '16 at 23:22
  • $\begingroup$ @Goats: in that case it can happen that both maps are isomorphisms. $\endgroup$ – Qiaochu Yuan Jan 18 '16 at 0:36

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.