Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $T$ be a linear transformation on $V$ over field $F$. Let $c\in F$ be a characteristic value of $T$, and $W_c$ the characteristic subspace of $T$ associated with $c$. Suppose a proper $T$-invariant subspace $W \supset W_c$, and there is a vector $\alpha\in V\setminus W$, such that $(T-cI)\alpha \in W$. Prove that the minimal polynomial $p_T$ of $T$ is in the form of $(x-c)^2q$ for nonzero polynomial $q\in F[x]$

My progress:

My first thought is to use $p_\alpha\mid p_T$, and I need to show $(T-cI)^2\alpha=0$. for a certain vector $\alpha$ in $V\setminus W$, $(T-cI)\alpha$ is not zero because otherwise $\alpha$ is in $W_c$ and thus is in $W$ (contradiction).

So if I can show $W=W_c$, then it's done. But I fail to show $W=W_c$. Or am I in the wrong way?

share|cite|improve this question
Out of curiosity, why delete and repost? – Jonas Meyer Aug 1 '12 at 1:50
I corrected some writting, it didnot get answer for the day – David Aug 1 '12 at 3:30
To correct writing, editing seems to work well. There is no need to delete and repost. This thread on meta might be relevant. – Jonas Meyer Aug 1 '12 at 3:34
What is $p_\alpha$? $W$ does not have to equal $W_c$. – Jonas Meyer Aug 1 '12 at 3:39
$p_\alpha$ is the T-annihilator associated to $\alpha$, which divides $p_T$. can you offer me some hints? – David Aug 1 '12 at 17:02

You should start realising how little you are asked to do: it is given that $c$ is an eigenvalue (characteristic value), so a root of$~p_T$, and the (monic) factor $q$ can absorb anything; therefore all that is asked is to show that $c$ is a multiple root of the minimal polynomial $p_T$. Although I initially thought that "characteristic subspace" meant "generalized eigenspace" (blame my French connection, since that is precisely what "sous espace caractéristique" means), I will take it that $W_c$ here is just the ordinary eigenspace for$~c$ of$~T$, as with the other interpretation the hypothesis of question becomes absurd as we shall see.

It is given that for some $T$-stable subspace $W$ containing the characteristic subspace$~W_c$, there exists $\alpha\in V\setminus W$ with $(T-cI)v\in W$. This means that the image of $\alpha$ in the quotient space $V/W$ is an eigenvector with eienvalue$~c$ for the linear map $T_{/W}$ induced by $T$ in $V/W$. The existence of such a vector shows that the algebraic multiplicity of$~c$ as eigenvalue of$~T$ (on all of $V$) is strictly larger than its algebraic multiplicity as eigenvalue of the restriction$~T|_W$ of$~T$ to$~W$. This is because the characteristic polynomial of$~T$ is the product of the characteristic polynomials of $T|_W$ and of $T_{/W}$, the latter having$~c$ as root. But since $W$ already contains the whole eigenspace $W_c$, the geometric multiplicities of$~c$ as eigenvalue of $T|_W$ and of $T$ are the same. Hence the generalised eigenspace of$~T$ cannot be$~W_c$ (and in particular $T$ cannot be diagonalisable), nor indeed can it be contained in$~W$. Then there is some vector $v\in V$ with $(T-cI)^2(v)=0$ but $(T-cI)(v)\neq0$, which shows that $c$ is indeed a multiple root of$~p_T$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.