0
$\begingroup$

Let $K$ be a field and $E$ be a $K$-vector space of dimension $n$. Let $\phi$ be an endomorphism of $E$.

Let $(\lambda_1,\cdots,\lambda_n)$ be a family of distinct scalars and $(x_1,\cdots,x_n)$ be a family of vectors such that $\phi(x_i)=\lambda_ix_i$ for each $i$.

I want to show that the family $(x_1,\cdots,x_n)$ is linearly independent. I have two clear hints

1) for each positive integer $k$, $\phi^k(x_i)=\lambda_i^kx_i$

2) If there exists a family of scalars $(\alpha_1,\cdots,\alpha_n)$ such that $\sum_{i=1}^{n}\alpha_ix_i=0$ then for each positive integer $k$, $\sum_{i=1}^{n}\alpha_i\lambda_i^kx_i=0$

I can see easily the two hints but I don't see how to proceed to show that the family of vectors is linearly independent. Possibly with the second hint if we set the equation $\sum_{i=1}^{n}\alpha_ix_i=0$ and want to show that all the scalars $\alpha_i$ must be zero but how to do that, thank you for your help!

$\endgroup$
1
$\begingroup$

Another hint:

You have to suppose the vectors $x_i$ are non-zero.

Consider the linear map from $E^n$ to $E^n$ which maps $\begin{bmatrix}x_1\\\dots\\x_n\end{bmatrix}$ to $\begin{bmatrix}\sum_{i=1}^n\alpha_i x_i\\\phi(\sum_{i=1}^n\alpha_i x_i)\\\vdots\\\phi^{n-1}(\sum_{i=1}^n\alpha_i x_i)\end{bmatrix}$.

Write the matrix of the system obtained for $k=1,2,\dots,n$. What is its determinant?

$\endgroup$
2
  • $\begingroup$ The matrix in which basis ? remember that the $x_i$ are vectors not scalars so turning the hint 2 into a matrix does not seem to be a valid operation. I think you want to compute the determinant of $(x_1, \cdots,x_n)$ in some basis and show that it is not zero, is it true ? what would be that basis ? $\endgroup$
    – palio
    May 5 '16 at 18:37
  • $\begingroup$ @palio: I've added some details. Matrices can also operate on systems of vectors. $\endgroup$
    – Bernard
    May 5 '16 at 18:48

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.