Consider the linear transformation from Cn into itself that sends a vector x ∈ ker(A−λj I )αj to λj x (j = 1, . . . , m). Let L be the n×n matrix associated with this linear transformation. This implies Lx = λj x for all x ∈ ker(A − λj I )αj . Clearly, ker(L − λj I ) = ker(A − λj I )αj for j = 1, . . . , m. Therefore, there exists a basis of Cn that consists of eigenvectors of L. Consequently, L
is diagonalizable.