-1
$\begingroup$

I read a very slick proof of determinant properties, in this case of the fact $\det A = \det A^T$, which says in one place

It suffices to notice that for any elementary matrix $M$ we have $\det M = \det M^T$

But it's not obvious to me. How can we show it?

Note: we're using the Laplace expansion for rows as the definition. It was noted that for elementary matrix $M$ we have $\det M = 1$ if $M$ adds a multiplied row, $\det M = -1$ if $M$ swaps two rows and $\det M = c$ if $M$ multiplies one row by $c$.

$\endgroup$
  • 1
    $\begingroup$ Try writing each elementary operation as a matrix, then take its determinant and transpose. $\endgroup$ – vadim123 Feb 9 '16 at 21:52
3
$\begingroup$

We have 3 cases:

1) $M$ multiplies a row by $c$

Then $M$ looks like $$ \left[ \begin{matrix} 1 & & \\ & 1 \\ & & \ddots \\ & & & 1 \\ & & & & c \\ & & & & & 1 \\ & & & &&&\ddots \\ & & & &&&& 1 \end{matrix} \right] $$ So trivally $M^T = M$, hence the equality of deteminants

2) $M$ adds another row. Then the $M$ looks like $$ \left[ \begin{matrix} 1 & & \\ & 1 & \ldots & a \\ & & \ddots & \vdots \\ & & & 1 \\ & & & &\ddots \\ & & & && 1 \end{matrix} \right] $$ By doing the Laplace expansion at the column containing $a$, we get $$detM = (-1)^{i+j} a \det I + (-1)^{j+j} \det I$$. This is symmetrical wrt to $i,j$, so in this case $\det M = \det M^T$ too

3) $M$ swaps two rows. Then $M$ looks like $$ \left[ \begin{matrix} 1 & & \\ & \ddots \\ & & 1 \\ & & & 0 & & & & 1 \\ & & & & 1 \\ & & && & \ddots \\ & & & & & & 1 \\ & & & 1 & & & & 0 \\ & & & & & & & & 1 \\ & & & & & & & & & \ddots \\ & & & & & &&& &&1 \end{matrix} \right] $$ This matrix is symmetrical, so $M = M^T$, hence the equality of determinants.

$\endgroup$
  • $\begingroup$ Your second case is wrong, In (2) it's not necessarily the case that $\tilde M_{ij}=I$. $\endgroup$ – Postal Model May 8 '18 at 8:12
  • $\begingroup$ Say $M=I_n$, then $\tilde M_{ij}=I_{n-1}$ iff $i=j$. And albeit you're wrong at this part your result is correct because the $\det(\tilde{I_{ij}})=0$ when $i\not=j$ so $(-1)^{i+j} a \det(\tilde{I_{ij}})$ is zero. $\endgroup$ – Postal Model May 8 '18 at 8:26

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.