I'm reading Axler Sheldon's paper Down with Determinants! and on page 7 he states the following:

4. The Minimal Polynomial

Because the space of linear operator on $V$ is finite dimensional, there is a smallest positive integer $k$ such that

$$I, T, T^2, ..., T^k$$

are not linearly independent.

In above $T$ is a linear operator on $V$ and $V$ is a $n$-dimensional complex vector space. The author doesn't seem to explain why the space of linear operators on $V$ is finite. He simply seems to states this.

Question: Why is it finite dimensional? Is this a consequence from something?

Subquestion: The author also uses on some places of the paper the notation $T|_{V_2}$ where $V_2$ is a subspace of V. What does this notation mean? Is it the restriction of $T$ to $V_2$?

  • $\begingroup$ subanswer: yes, it is the restriction $\endgroup$ – Bananach Aug 11 '16 at 10:26
  • $\begingroup$ Thank you appreciate it! =) $\endgroup$ – jjepsuomi Aug 11 '16 at 10:28

Choose a basis of $V$, say $e_1, \dots, e_n$. Any linear operator $T$ on $V$ is determined uniquely by its action on the basis, i.e. by the values $$T(e_i) = a_{i1} e_1 + \dots + a_{in} e_n, \quad 1 \le i \le n,$$ for some $a_{ij} \in \mathbb{C}$. In other words, $T$ is uniquely determined by the $n^2$ complex numbers $a_{ij}$. This implies that the space of operators has dimension $n^2$, and so is finite-dimensional. Using your basis to view the operators as matrices is another easy way to see the space is of dimension $n \times n = n^2$.

| cite | improve this answer | |
  • $\begingroup$ Understood :) Thank you! $\endgroup$ – jjepsuomi Aug 11 '16 at 10:28

That is because, over any field $K$, in the finite dimensional vector space $V$ choose a basis $\mathcal B=(e_1,\dots,e_n)$ of $V$. Then, for any vector space $W$, we have a (non-canonical) isomorphism: $$\operatorname{Hom}_K(V,W)\cong W^{\lvert\mathcal B\rvert}=W^n. $$ In particular $$\operatorname{End_K}(V)=\operatorname{Hom_K}(V,V)\cong V^n.$$


Actually, this theorem is more generally true for any commutative ring $R$ and any finitely generated free $R$-module.

| cite | improve this answer | |
  • $\begingroup$ Thank you very much! Appreciate it =) $\endgroup$ – jjepsuomi Aug 11 '16 at 10:29
  • $\begingroup$ that explains why the space of operators from $V$ to $\mathbb{R}$ is finite-dimensional. The question is about operators from $V$ to $V$. $\endgroup$ – Bananach Aug 11 '16 at 10:38
  • $\begingroup$ @Bananach: I misundersood the question. Thanks for pointing it. I'll change my answer. $\endgroup$ – Bernard Aug 11 '16 at 10:41
  • $\begingroup$ @Bananach: Done. $\endgroup$ – Bernard Aug 11 '16 at 10:49

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.