Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Given an $n\times n$ matrix A with real entries such that $A^2=-I$, prove (a) that $n$ is even and (b) that $A$ has no real eigenvalues. How do you do this? I have no idea where to start.

share|cite|improve this question
For the first part, consider the determinant. For the second, consider what happens to an eigenvector with a real eigenvalue if you apply $A^2$ to it. – Tobias Kildetoft Aug 19 '13 at 18:45
Actually (b) implies (a), since every odd-dimensional algebraic equation has a real root. – njguliyev Aug 19 '13 at 18:52
up vote 18 down vote accepted


Since the matrix has real entries, $\det A$ is real, so $\det A^2 = (\det A)^2$ is positive. But $\det -I = (-1)^n$, because we can just multiply down the main diagonal, so we must have that $n$ is even.


Suppose $A$ has a real eigenvalue $\lambda$ with corresponding eigenvector $v$. Then $$-v=-Iv=A^2v=A(Av)=A(\lambda v)=\lambda^2v.$$

This products a contradiction, because no real number squares to $-1$.

share|cite|improve this answer

HINT: $\det(-I)=\det(A^2)$. If $Ax=\lambda x$, then $-x=A^2x=\ldots\;$?

share|cite|improve this answer

Take a look at Cayley Hamilton Theorem. An overkill but the two people above me already answered it! In particular, look at the minimal polynomial.

share|cite|improve this answer

Let's start with (b). Suppose $\lambda$ were a real eigenvalue of $A$. Then there would exist a nonzero vector $v \in R^n$ such that

$Av = \lambda v$.


$A^2v = \lambda^2 v$,


$(A^2 + I)v = (\lambda^2 + 1)v$;

but since $A^2 + I = 0$ this implies

$(\lambda^2 + 1)v =0$,

which since $v \ne 0$ implies

$\lambda^2 + 1 = 0$;

but no real $\lambda$ satisfies this equation. This contradiction shows $A$ has no real eigenvalues.

As for (a), if $n$ were odd, then the characteristic polynomial $\det(A - \lambda I)$ of $A$ would have odd degree $n$; but every polynomial with real coefficients and odd degree has at least one real root. But we have seen that $A$ can have no real eigenvalues. Thus $n$ must itself be even. QED.

Hope this helps. Cheers.

share|cite|improve this answer

To prove (a):

Consider the matrix $I_n$, the $n*n$ identity matrix. Clearly, $\det I_n = (-1)^n$. Now, consider $\det A$. As $A$ has only real entries, $\det A$ is real. Thus, $(\det A)^2$ is positive. Since $A^2 = I_n$, it is also the case that $(\det A)^2 = \det I_n = (-1)^n$, and as $(\det A)^2 > 0$, it must be the case that $\det I_n > 0$, so $n$ must be even.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.