5
$\begingroup$

Why does $\det(\text{adj}(A)) = 0$ if $\det(A) = 0$? (without using the formula $\det(\text{adj}(A)) = \det(A)^{n-1}.)$

$\endgroup$
1
  • $\begingroup$ Note that you need $A \in \mathbf{k}^{n\times n}$ with $n \geq 2$ here. $\endgroup$ Commented Jul 29, 2014 at 15:59

2 Answers 2

11
$\begingroup$

Use : $$\mathrm{adj}(A)\times A=A \times \mathrm{adj}(A)=\mathrm{det}(A)\mathrm{I}_{n}.$$

Let's assume that $\mathrm{det}(A)=0$. If $A=0$, then $\mathrm{adj}(A)=0$ as well and, obviously, $\mathrm{det}\big( \mathrm{adj}(A) \big)=0$. Otherwise, there exist $x \in \mathbb{R}^{n}$, $x \neq 0$ such that $Ax \neq 0$. It follows that $\mathrm{adj}(A) Ax=0$. As a consequence, $\mathrm{Span}(Ax) \subset \mathrm{ker} \; \mathrm{adj}(A)$ (which means that $\mathrm{ker} \; \mathrm{adj}(A) \neq \lbrace 0 \rbrace$). Therefore, $\mathrm{adj}(A)$ is not invertible and $\mathrm{det}\big( \mathrm{adj}(A) \big)=0$.

$\endgroup$
0
0
$\begingroup$

In that case, shall we use the following? $[adj(A)]^{-1} = adj(A^{-1})$ ?

Assume that $adj(A)$ is non-singular. Then inverse of adj(A) exists.

But we know $[adj(A)]^{-1}=adj(A^{-1})$ which is contradiction because it is given that $A$ is singular so $A^{-1}$ does not exist and hence adjoint of $A^{-1}$ also does not exist.

hence the proof.

$\endgroup$
4
  • 1
    $\begingroup$ The identity you are using ($\left(\operatorname{adj} A\right)^{-1} = \operatorname{adj}\left(A^{-1}\right)$) is usually proven for $A$ invertible. $\endgroup$ Commented Jul 29, 2014 at 15:58
  • $\begingroup$ @darijgrinberg I am sorry I could not get your point. Can you please explain me further more? I mean I was unable to understand if you are saying what I have shown is incorrect. Please clear my doubt $\endgroup$
    – KON3
    Commented Jul 29, 2014 at 16:02
  • $\begingroup$ I'm just saying that you are using a fact that, in the form you are using it, isn't standardly known. $\endgroup$ Commented Jul 29, 2014 at 16:07
  • $\begingroup$ Ohh ok that one. Ya, may be true, but why can't we solve a particular problem in several ways we want? After all, the answer has been given already. I just played another way. Anyway, thanks for the information. $\endgroup$
    – KON3
    Commented Jul 29, 2014 at 16:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .