I'm working my way through Franklin's Matrix Theory and am stuck on question 1.3.2.

The question:

Let $A$ have the form:

$$ A = \begin{bmatrix} a_{11} & a_{12} & a_{13} & a_{14} & a_{15}\\ a_{21} & a_{22} & a_{23} & a_{24} & a_{25}\\ 0 & 0 & a_{33} & a_{34} & a_{35}\\ 0 & 0 & a_{43} & a_{44} & a_{45}\\ 0 & 0 & a_{53} & a_{54} & a_{55} \end{bmatrix} = \begin{bmatrix} A_1 & \cdots \\ \bigcirc & A_2 \end{bmatrix} $$

Show that:

$$\det{A} = \det{A_1} \cdot \det{A_2}$$

The only argument I'm confident enough in making is that $\det{A}$ will ultimately consist only of elements from $A_1$ and $A_2$.

My reasoning is this:

Each term $s(j_1,\cdots,j_n)\prod_{i=1}^{n}a_{i j_i} $ of $\det{A}$ is non-zero only if whenever $1 \leq i \leq 2$, $1 \leq j_i \leq 2$ and whenever $3 \leq i \leq 5$, $3 \leq j_i \leq 5$.

This is because if when $1\leq i \leq 2$ we chose $j_i$ such that $3 \leq j_i \leq 5$ we would unavoidably have to choose a zero when $3 \leq i \leq 5$ as we must make one choice for each row 3,4,5. We want non-zero terms so for each choice of $j_i$ we pick from columns 3,4,5. However, one of the columns that would give us an element in $A_2$ (non-zero) is no longer available due to our previous choice, forcing us to take a zero.

The case which arises when $3 \leq i \leq 5$ and we choose $1 \leq j_i \leq 2$ is obvious.

I'm not confident in asserting anything beyond that.

Any help is much appreciated!

  • $\begingroup$ Consider polylinear , skew-symmetric function of $n$ vectors. Then $F(v_{1} \dots v_{n}) = F(e_{1}\dots e_{n})detA$, where $A$ matrix of coordinates $v_{1} \dots$ in basis $e_{1} \dots$ $\endgroup$
    – openspace
    Feb 21 '17 at 7:46
  • $\begingroup$ Thank you, unfortunately I can't see the connection though. Probably just because I'm not familiar enough with polylinear functions or skew-symmetry. $\endgroup$
    – Casey
    Feb 21 '17 at 8:21


Here you can see the proof for a matrix of the form $$ \begin{bmatrix}A&0\\C&D \end{bmatrix} $$

with a little effort you can modify it for your case.

  • $\begingroup$ Ah, it's pretty obvious then in the matrix decomposition case I suppose. Following from the determinant property $\det{AB} = \det{A}\det{B}$. Now I'm going to try and learn more about a proof involving the Leibniz definition of the determinant. Thank you! $\endgroup$
    – Casey
    Feb 21 '17 at 8:18

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.