I am asked the following:

Assume that $M$ is skew-symmetric and invertible. If $M$ is an $n \times n$ matrix, prove that $n$ is even.

The following is my attempt at the solution. I'm not too sure if it is right, or if I'm allowed to do this.

So, if a matrix is skew-symmetric, then,


and by taking the determinant of both sides, $$\det(A^T)=\det(-A)$$

Using the property of transpose and the multi-linear property,


Here is where my problem arises. Are we allowed to assume that $n$ is even and then show that it is invertible? However, if we do this we get

$$\det(A) = \det(A)$$

and I'm not too sure what this equality means. Or where to go from here. Any help would be greatly appreciated.

  • 1
    $\begingroup$ Why are you trying to prove the converse of the claim? The converse is not true. If $n$ is even, $A$ may and may not be invertible. $\endgroup$ – darij grinberg May 26 at 10:07
  • $\begingroup$ What are you suggesting I do? I've been a bit stuck $\endgroup$ – Joshie Bread May 26 at 10:14
  • 1
    $\begingroup$ Nothing; you're done. You have to prove that $n$ is even if $A$ is invertible. You have shown that if $n$ is odd, then $A$ is not invertible. This is an equivalent statement. $\endgroup$ – darij grinberg May 26 at 10:21
  • $\begingroup$ Great! Thank you/. $\endgroup$ – Joshie Bread May 26 at 10:23

Note that if $n$ is odd, you would get $\det (A) = - \det (A)$, so A has determinant $0$ and is not invertible. Some invertible matrices that are even by even are skew-symmetric; try constructing them yourself!

  • $\begingroup$ Thank you. Is what I have done valid though for the proof? I am trying to prove that It is has an even $n$, Is it fair to assume that it is even? $\endgroup$ – Joshie Bread May 26 at 10:08
  • $\begingroup$ @JoshuaPeters Not quite; to prove $n$ is even you will need to prove that odd $n$ fails as in my answer. $\endgroup$ – auscrypt May 26 at 10:10
  • $\begingroup$ So a proof by contradiction? where we assume that $n$ is odd and invertible. Then we arrive at a conclusion where $2 det(M)=0$ which implies it is not invertible. So $n$ must be even. $\endgroup$ – Joshie Bread May 26 at 10:13
  • $\begingroup$ @JoshuaPeters Yes, that’s right. $\endgroup$ – auscrypt May 26 at 10:15
  • $\begingroup$ Is it fair to say that since the negation of the statement is false, this implies that if n is even, M is skew-symmetric and invertible. $\endgroup$ – Joshie Bread May 26 at 10:53

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.