I am reading a paper about operator theory, there is a proposition I could not understand.

Let $T\in L(X)$ be a quasi-nilpotent operator and $X_{1}$ be a non-zero finite-dimensional subspace of X, then the restriction of T to $X_{1}$ is nilpotent (there exist a positive integer n such that $T^{n}=0$)?

Is this correct?

  • $\begingroup$ Yes, it is correct. $\endgroup$ – scineram Sep 16 '13 at 16:03
  • 1
    $\begingroup$ @scineram why is it true? $\endgroup$ – Norbert Sep 16 '13 at 16:32
  • 3
    $\begingroup$ Is $X_1$ an invariant space of $T$? $\endgroup$ – Jonathan Y. Sep 16 '13 at 19:11
  • $\begingroup$ Do you have a link to the paper? $\endgroup$ – Davide Giraudo Sep 16 '13 at 19:19
  • 1
    $\begingroup$ I think there should be other assumptions, otherwise Volterra's operator en.wikipedia.org/wiki/Nilpotent_operator seems to give a counter-example. $\endgroup$ – Davide Giraudo Sep 16 '13 at 20:33

I'll assume $X_1$ is an invariant subspace of $T$. In that case, we know for all $\lambda\neq 0$ that $(T|_{X_1}-\lambda I_{X_1})$ is injective, and, since $X_1$ is finite-dimensional, invertible. Therefore $\sigma(T|_{X_1})\subset\{0\}$. It's also non-empty, so $T|_{X_1}$ is also quasinilpotent. Finally, as is apparent from the Jordan form of matrices, every quasinilpotent matrix is nilpotent.


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.