Given an $n \times n$ matrix $A$ with real entries such that $A^2=-I$. Prove that A has no real eigenvalues.

We can easily prove the following additional statements about $A$ by taking determinants of both sides of the given equality -

(a) $A$ is nonsingular;

(b)$n$ is even;



Well, what does it mean for something to be a real eigenvalue of A? It means there's some vector we can multiply into A on the right and get back a scalar multiple of that vector.

But for any eigenvector, if we do this twice: $-v = (-I)v = A^2 v = A(A v) = A(\lambda v) = \lambda^2 v$.

You should be able to take it from here.


Suppose $\lambda$ were a real eigenvalue of $A$; then there is a non-zero vector $\vec v \in \Bbb R^n$ such that $A\vec v = \lambda \vec v$. This implies that $A^2 \vec v = A(A\vec v) = A(\lambda \vec v) = \lambda (A \vec v) = \lambda^2 \vec v$; then $0 = (A^2 + I) \vec v = (\lambda^2 + 1) \vec v$; since $\vec v \ne 0$, we have $\lambda^2 + 1 = 0$, clearly impossible for real $\lambda$ ; thus $A$ has no real eigenvalues. This fact implies that $n$ is even, since otherwise the characteristic polynomial $p_A(\lambda) = \det(A - \lambda I)$ of $A$ always has at least one real root by virtue of the fact that $\deg p_A(\lambda) = n$ is odd. Furthermore, since $A^2 = -I$ and $n$ is even, we have $(\det A)^2 = \det A^2 = \det(-I) = 1$. Thus $\det A \ne 0$ so $A$ is nonsingular. Finally $A^2 + I = 0$ actually implies $\lambda^2 = -1$ for the complex eigenvalues of $A$, by a very slight extension of our previous argument, thus $\lambda = \pm i$; but $A$ real implies the $\lambda$ occur in complex conjugate pairs, so that $\det A$, being the product of $n / 2$ pairs $\lambda, \bar \lambda$ with $\lambda \bar \lambda = i(-i) = 1$, must itself be $1$. QED.

And that cover's all the OP's points, if I am not mistaken.

Hope this helps! Cheerio,

and as always,

Fiat Lux!!!


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.