Theorem 1. Let A be an n n matrix, and let v1 , v2 , . . . , vn be (nonzero) eigenvectors for A with distinct eigenvalues.

. Then these vectors are linearly independent. Proof. Assume theyre linearly dependent. By reordering the vi s if necessary, there exists some k such that 1 k n and coecients a1 , a2 , . . . , ak such that
k

0=
i=1

ai vi

and ai = 0 for all i. Let i be the eigenvalue associated to vi for each i. Then 0 = (A k I )0
k

= (A k I )
i=1 k

ai vi

=
i=1 k

ai (A k I )vi ai (Avi k I vi )
i=1 k

=
i=1 k

ai (i vi k vi ) ai (i k )vi
i=1 k1

=
i=1

ai (i k )vi .

For all i < k , i k = 0, since the vi s were presumed to have distinct eigenvalues. Therefore ai (i k ) = 0 for all i < k . Now we have obtained a linear dependence relation
k1

0=
i=1

ai (i k )vi

involving only the vectors v1 , v2 , . . . , vk1 where all of the coecients are nonzero. We can repeat this process (the next step is to multiply by the matrix A k1 I , etc.) until we arrive at an equation 0 = v1 for some = 0. This implies v1 = 0, which is a contradiction. Therefore v1 , v2 , . . . , vn are linearly independent.