Properties of 2x2 Determinants

If you treat the determinant only as the formula $ad - bc$, half its power is hidden. The real magic is in the properties: rules that let you shortcut huge calculations, deduce facts about a matrix without expanding, and connect algebra to geometry. This section catalogues the five most useful properties for the $2 \times 2$ case, with the same parallelogram picture as your guide.

The five fundamental properties

1. Multiplicativity: $\det(AB) = \det(A) \det(B)$.
2. Transpose: $\det(A^T) = \det(A)$.
3. Row swap: swapping two rows multiplies the determinant by $-1$.
4. Row scaling: multiplying a row by $c$ multiplies the determinant by $c$.
5. Row addition: adding a multiple of one row to another leaves the determinant unchanged.

Every one of these can be verified by direct expansion of $ad - bc$, but the geometric picture makes them inevitable. Each row of a $2 \times 2$ matrix is one side of the parallelogram. Swapping the rows flips the orientation, hence the sign change. Scaling a row stretches one side, the area scales linearly. Adding a multiple of one row to another shears the parallelogram, which preserves area. These are not arbitrary rules; they are geometry.

• Numerical Computing: Gaussian elimination relies on the fact that row operations either leave the determinant invariant or scale it predictably; LU decomposition uses this to compute determinants in $O(n^3)$ instead of the $O(n!)$ of cofactor expansion.
• Linear Programming: The simplex algorithm moves between corner points of the feasible region; each move is a row operation, and the determinant's sign tells you whether the move preserves orientation.
• Robotics: Singular configurations of a robot arm, where it loses a degree of freedom, are detected by checking when the Jacobian's determinant vanishes; controllers monitor this in real-time to avoid lock-up.

(Use the widget to see properties 3–5 in action. Drag $\mathbf{u}$ and $\mathbf{v}$ to fix $\det = 10$. Now imagine swapping the labels, the sign flips to $-10$. Imagine doubling $\mathbf{u}$: the parallelogram is twice as tall, area is $20$. Imagine adding $\mathbf{v}$ to $\mathbf{u}$: the parallelogram shears, but its area stays at $10$.)

Two consequences that come out for free

From the product rule (#1), applied to $A \cdot A^{-1} = I$:

$\det(A) \det(A^{-1}) = \det(I) = 1 \quad \Longrightarrow \quad \det(A^{-1}) = \frac{1}{\det(A)}.$

So the inverse matrix scales area by exactly the reciprocal of what $A$ scales by. Makes sense: undoing a $5\times$ stretch should mean a $\tfrac{1}{5}\times$ shrink.

From the row scaling property applied to both rows at once:

$\det(cA) = c^2 \det(A) \quad \text{for any } 2 \times 2 \text{ matrix.}$

More generally, $\det(cA) = c^n \det(A)$ for an $n \times n$ matrix, one factor of $c$ per row.

The dependence consequence

If one row of $A$ is a scalar multiple of the other (say row 2 = $k \cdot$ row 1), then the rows are linearly dependent. They point along the same line, so the parallelogram collapses to a line segment with area $0$. Therefore $\det(A) = 0$, and $A$ is singular. This is the cleanest way to spot a non-invertible matrix at sight: do any two rows look proportional?

Row reduction as determinant arithmetic

Properties 3, 4, 5 are exactly the three elementary row operations from Gaussian elimination. So they let you compute $\det(A)$ by reducing $A$ to a triangular form, tracking sign changes from swaps and scale factors from row scalings, and then just multiplying the diagonal entries. This is how a computer evaluates large determinants: row operations, not $ad - bc$ generalisations.

Try it

• If $\det(A) = 4$ and $\det(B) = -3$, compute $\det(AB)$, $\det(A^{-1})$, $\det(2A)$, and $\det(A^T)$.
• Without expansion, find \det\begin{pmatrix} 1 & 2 \\ 3 & 6 \end{pmatrix}. (Hint: look at the rows.)
• Start from \det\begin{pmatrix} 2 & 5 \\ 1 & 3 \end{pmatrix}. Verify the row-addition property: subtract $\tfrac{1}{2}$ times row 1 from row 2, then compute $\det$ of the new matrix, you should get the same answer.
• Use the widget to see why $\det(cA) = c^2 \det(A)$: double $\mathbf{u}$ and double $\mathbf{v}$. The area quadruples, not just doubles.

Pause: property 5 says adding row 1 to row 2 leaves $\det$ unchanged. But property 4 says scaling row 2 by $2$ multiplies $\det$ by $2$. Adding row 1 to row 2 and then scaling that new row by $2$, what happens to $\det$? Apply the two rules in order.

A trap to watch for

Property 1, $\det(AB) = \det(A) \det(B)$, is the famous one. Beginners over-apply it: they write $\det(A + B) = \det(A) + \det(B)$. This is false in general. Test it: $A = B = I_2$ gives $\det(A + B) = \det(2I_2) = 4$, but $\det(A) + \det(B) = 1 + 1 = 2$. The determinant is multiplicative in products, but it is not linear in sums. Memorise the asymmetry.