15
$\begingroup$

As the title says, considering the roots of the characteristic polynomial are the eigenvalues of the matrix, and in the complex field every polynomial has $n$ different roots where $n$ is the polynomial degree, this means every eigenvalue on the complex field is a simple eigenvalue, so every matrix in $\mathbb C$ will be diagonalizable. Am I wrong?

$\endgroup$
3
  • $\begingroup$ What is an autovalue? $\endgroup$
    – Bernard
    Commented Sep 14, 2016 at 18:13
  • 2
    $\begingroup$ I assume "autovalue" means eigenvalue $\endgroup$ Commented Sep 14, 2016 at 18:21
  • $\begingroup$ How did you get that every polynomial has $n$ different roots where $n$ is the degree of the polynomial in complex field ?? $\endgroup$ Commented Sep 20, 2018 at 5:40

1 Answer 1

30
$\begingroup$

No, not every matrix over $\Bbb C$ is diagonalizable. Indeed, the standard example $\begin{pmatrix} 0&1\\0&0 \end{pmatrix}$ remains non-diagonalizable over the complex numbers.

Here's where your argument breaks down. You've correctly argued that every $n\times n$ matrix over $\Bbb C$ has $n$ eigenvalues counting multiplicity. In other words, the algebraic multiplicities of the eigenvalues add to $n$. However, the geometric multiplicities are not necessarily the same as the algebraic multiplicities (when the algebraic multiplicity is $k$ then the geometric multiplicity can be any integer from $1$ to $k$). And diagonalizability is equivalent to the sum of the geometric multiplicities being $n$, not the sum of the algebraic multiplicities.

Looking at your exact wording, you have said "every polynomials have $n$ different roots"; that's a mistake—complex polynomials of degree $n$ have $n$ roots counting multiplicity, but nothing stops those roots from coinciding sometimes—consider $(x-i)^n$ for example.

$\endgroup$
3
  • 1
    $\begingroup$ Oh, got it... ...i do so much confusion... $\endgroup$ Commented Sep 14, 2016 at 18:50
  • $\begingroup$ @Greg Martin How do we know that diagonalizability is equivalent to the sum of the geometric multiplicities being n? $\endgroup$ Commented Mar 29, 2021 at 4:15
  • $\begingroup$ the geometric multiplicities are the dimensions of the eigenspaces, if they add up to the dimension of the full space, we have an eigenbasis, with respect to which the matrix is diagonal. $\endgroup$
    – peter
    Commented May 3, 2021 at 4:03

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .