4

$T$ is a linear operator above field $F$. V is of finite dimension and $dim Im T = 1$; show that $T$ is diagonalizable or nilpotent but not both.

This seems like a rather strong statement. I know that $T$ is diagonalizable if there is a basis $B$ of eigenvectors, also from $dimImT = 1$ I think that there are at most 2 eigenvalues because $ImT$ is a span of a single vector, therefore there is a vector $v\in V\;$ such that $T(v) = \lambda v$.

I am kind of missing the rail of the proof here. I would like to get some insight / help.

Thanks in advance!

GoodWilly
  • 915

1 Answers1

3

Say $Im T$ is spanned by $v\ne0$. Choose $w$ so $v=Tw$. And choose a scalar $\lambda$ so $Tv=\lambda v$.

If $\lambda=0$ then $Tv=0$; hence $T^2=0$, so $T$ is nilpotent. Further, $T^2=0$, $T\ne0$ imply that $T$ is not diagonalizable.

Suppose $\lambda\ne0$. Then $T^kv=\lambda^kv\ne0$. So $T^k\ne0$, which is to say $T$ is not nilpotent. And if $n=\dim(V)$ and $Z$ is the nullspace of $T$ then $dim(Z)=n-1$. Say $b_1,\dots,b_{n-1}$ is a basis for $Z$. Let $b_n=v$. Then the set $B=\{b_1,\dots,b_n\}$ is an independent set of $n$ vectors. So $B$ is a basis for $V$; since every vector in $B$ is an eigenvector of $T$ this shows that $T$ is diagonalizable.