Let $V$ be a finite-dimensional vector space and let $T : V \to V$ be a linear transformation. Let $U, W\subseteq V$ be $T$-invariant subspaces of $V$, such that $U\oplus W=V$. Let $q\in \mathbb{F}[x]$ be a polynomial. Given that $U$ is a sub-space of $\mbox{im }q(T)$, show that $\ker q(T)$ is a sub-space of $W$.

I tried to show the $U \cap \ker q(T)= \{0\}$. Also, solving directly for the definitions. I also know that every $T$-invariant subspace is also $q(T)$-invariant. Any help would be very helpful.


Since $q(T)|_U$ is surjective and $U$ is finite-dimensional, $q(T)|_U$ is bijective. Therefore, its kernel is $\{0\}$.

So, if $v\in\ker q(T)$, write $v$ as $u+w$, with $u\in U$ and $w\in W$. Now, $$T(u+w)=0\iff q(T)(u)=q(T)(-w).$$But $q(T)(-w)\in W$ (because $W$ is invariant) and $q(T)(-w)\in U$ (because $\operatorname{im}q(T)\subset U$). Therefore, $q(T)(-w)=0$, and so $q(T)(u)=0$. But $q(T)(u)=0\implies u=0$. Therefore, $v=0+w=w\in W$.

  • $\begingroup$ Oops. forgot to mention that $V$ is of finite dimension. Apologies. $\endgroup$ – user558964 May 4 '18 at 21:04
  • $\begingroup$ Though, thank you very much. This is still helpful. $\endgroup$ – user558964 May 4 '18 at 21:05
  • $\begingroup$ @user558964 I've edited my answer. $\endgroup$ – José Carlos Santos May 4 '18 at 21:10
  • $\begingroup$ I am missing one thing, why is $q(T)$ surjective? $\endgroup$ – user558964 May 4 '18 at 21:21
  • $\begingroup$ I see why $q(T)(W) \subseteq W$ but why does $q(T)(W) = W$? $\endgroup$ – user558964 May 4 '18 at 21:23

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.