Area and collinearity

Give three points on a graph paper and ask a computer to check whether they lie on one straight line, or to find the area of the triangle they make — and the machine does not measure with a ruler. It builds a $3\times3$ determinant out of the coordinates and reads the answer straight off the number. This single tool, drawn from the Section A "Algebra (Matrices and Determinants)" unit of ISC Class 12 Mathematics, does two jobs at once: a non-zero value gives the exact area of the triangle formed by the three points, and a value of exactly zero certifies that the points are collinear — that the "triangle" has collapsed onto a line. It is one of the cleanest applications of determinants in the whole syllabus, because it turns a geometric question (area, or straightness) into pure algebra: write down a grid of numbers, expand, and read off the answer.

The core insight

The signed determinant $\frac12\begin{vmatrix}x_1&y_1&1\\x_2&y_2&1\\x_3&y_3&1\end{vmatrix}$ gives the area of a triangle from its vertices, and equals zero exactly when the three points are collinear.

The core idea

A triangle is fixed once its three vertices are fixed, so its area must be some formula in the six coordinates $x_1,y_1,x_2,y_2,x_3,y_3$ . Writing each vertex as a row $(x_i,y_i,1)$ and taking the determinant of the resulting $3\times3$ matrix packages that formula compactly:

\Delta = \frac12\begin{vmatrix}x_1 & y_1 & 1\\ x_2 & y_2 & 1\\ x_3 & y_3 & 1\end{vmatrix}

where: $(x_1,y_1)$ , $(x_2,y_2)$ , $(x_3,y_3)$ are the triangle's vertices in the Cartesian plane, and $\Delta$ is the enclosed area. Depending on the order in which the vertices are listed (clockwise or anticlockwise), the raw expansion can come out negative — area itself is never negative, so the working rule is always to take the absolute value:

\Delta = \frac12\left|\begin{vmatrix}x_1 & y_1 & 1\\ x_2 & y_2 & 1\\ x_3 & y_3 & 1\end{vmatrix}\right|

Three points that do not form a genuine triangle — because they all sit on one line — must give an "area" of zero. That gives a second, equally important reading of the same expression: it is a built-in straightness test.

Collinearity test. Three distinct points $(x_1,y_1)$ , $(x_2,y_2)$ , $(x_3,y_3)$ are collinear if and only if

\begin{vmatrix}x_1 & y_1 & 1\\ x_2 & y_2 & 1\\ x_3 & y_3 & 1\end{vmatrix} = 0

with no $\tfrac12$ and no absolute value needed, since we are only asking whether the value is zero, not what it measures. This links straight back to the determinant properties earlier in the unit: a zero determinant means the three rows are linearly dependent. Geometrically, the rows here are the three points (each padded with a $1$ ), and linear dependence among them means the points fail to spread out into two independent directions — they are confined to one line. A non-zero determinant, by contrast, is a genuine measure of how far the points are from being collinear, scaled by $\tfrac12$ into an honest area.

Figure

Three labelled points on a coordinate grid joined into a triangle with the shoelace determinant written beside it, next to a second set of three points lying on one straight line with the same determinant expression equal to zero

The area-collinearity connection, derived

The goal is to show, not just assert, why the very same $3\times3$ determinant that gives an area for a genuine triangle collapses to zero the instant the three points fall on one straight line, using nothing beyond the row operations already established for evaluating determinants in this unit.

Step 1. Start from the general determinant $\begin{vmatrix}x_1 & y_1 & 1\\ x_2 & y_2 & 1\\ x_3 & y_3 & 1\end{vmatrix}$ and apply the row operations $R_2 \to R_2 - R_1$ and $R_3 \to R_3 - R_1$ . These operations leave the value of a determinant unchanged (a standard property from earlier in the unit), so the value is exactly the same as before, but the matrix is now $\begin{vmatrix}x_1 & y_1 & 1\\ x_2-x_1 & y_2-y_1 & 0\\ x_3-x_1 & y_3-y_1 & 0\end{vmatrix}$ .

Step 2. Expand this new form along the third column, which now has only one non-zero entry, the $1$ in row one. The cofactor of that entry is $\begin{vmatrix}x_2-x_1 & y_2-y_1\\ x_3-x_1 & y_3-y_1\end{vmatrix}$ , and every other term in the expansion vanishes because it is multiplied by a zero. So the original $3\times3$ determinant is exactly equal to this single $2\times2$ determinant, $(x_2-x_1)(y_3-y_1) - (y_2-y_1)(x_3-x_1)$ .

Step 3. Recognise what that $2\times2$ determinant means. Its two rows are the components of the direction vectors $\overrightarrow{P_1P_2} = (x_2-x_1,\, y_2-y_1)$ and $\overrightarrow{P_1P_3} = (x_3-x_1,\, y_3-y_1)$ , the arrows from the first vertex to the other two. A $2\times2$ determinant of two vectors is zero exactly when those vectors are parallel (one is a scalar multiple of the other) — and two arrows from the same starting point being parallel is precisely the geometric statement that $P_1$ , $P_2$ , $P_3$ all lie along one direction, i.e. are collinear.

Step 4. Conclude the two-way link. If the points are collinear, the direction vectors are parallel, the $2\times2$ determinant is zero, and by Step 2 the original $3\times3$ determinant is zero too — matching the stated collinearity test. If the points are not collinear, the direction vectors point in genuinely different directions, the $2\times2$ determinant is non-zero, and (with the $\tfrac12$ and absolute value applied) it measures the actual enclosed area. The area formula and the collinearity test are therefore the same calculation viewed two ways, not two separate formulas to memorise.

Worked example

Find the area of the triangle with vertices $A(1,2)$ , $B(4,6)$ , $C(5,2)$ , then use the same determinant machinery to test whether $P(2,3)$ , $Q(4,7)$ , $R(6,11)$ are collinear.

Step 1. Set up the area determinant for $A$ , $B$ , $C$ : $\Delta = \frac12\begin{vmatrix}1 & 2 & 1\\ 4 & 6 & 1\\ 5 & 2 & 1\end{vmatrix}$ .

Step 2. Expand along the first row: $\begin{vmatrix}1 & 2 & 1\\ 4 & 6 & 1\\ 5 & 2 & 1\end{vmatrix} = 1\begin{vmatrix}6&1\\2&1\end{vmatrix} - 2\begin{vmatrix}4&1\\5&1\end{vmatrix} + 1\begin{vmatrix}4&6\\5&2\end{vmatrix} = 1(4) - 2(-1) + 1(-22) = 4+2-22 = -16$ .

Step 3. Take half the absolute value: $\Delta = \frac12|-16| = 8$ . The area of triangle $ABC$ is 8 square units.

Step 4. Now test $P(2,3)$ , $Q(4,7)$ , $R(6,11)$ for collinearity: $\begin{vmatrix}2 & 3 & 1\\ 4 & 7 & 1\\ 6 & 11 & 1\end{vmatrix} = 2(-4) - 3(-2) + 1(2) = -8+6+2 = 0$ . The determinant is exactly zero, so $P$ , $Q$ , $R$ are collinear (each point indeed satisfies $y = 2x-1$ ).

Competitive stretch

Replace $R$ with $S(6,10)$ and repeat: $\begin{vmatrix}2 & 3 & 1\\ 4 & 7 & 1\\ 6 & 10 & 1\end{vmatrix} = 2(-3) - 3(-2) + 1(-2) = -6+6-2 = -2$ . Since this is $-2 \neq 0$ , $P$ , $Q$ , $S$ are not collinear; they form a genuine triangle of area $\tfrac12|-2| = 1$ square unit — a useful check that a small non-zero determinant, however tiny, still means "not collinear," never "almost collinear."

Common traps

Dropping the $\tfrac12$ factor, or forgetting to take the absolute value when the raw expansion comes out negative — area cannot be negative.
Mixing up column order: the standard layout is $x_i, y_i, 1$ across each row; swapping columns flips the sign of the determinant, which is harmless for area (absolute value fixes it) but confuses the collinearity test if done carelessly mid-calculation.
Treating a small but non-zero determinant as "nearly collinear" — the test is exact: only precisely zero counts, no matter how small the computed area is.
Reaching for this $3\times3$ determinant for points in three dimensions, or for vectors instead of points — this formula applies only to points in the Cartesian plane; 3-D and vector-cross-product area formulas belong to a different unit.

Ask the tutor: if you want more practice combining the area formula with finding a missing vertex coordinate, or a fresh collinearity check to try, ask the AI tutor about "area and collinearity."

Why do we take the absolute value in the area formula but not in the collinearity test?

The area formula is meant to output a genuine physical area, and area can never be negative, so once you have expanded the determinant you take $\tfrac12$ of its absolute value to get a sensible answer. The collinearity test asks a yes-or-no question — is the determinant exactly zero — and zero has no sign to worry about, so no absolute value is needed there. If you are testing collinearity, just expand the determinant and check whether the raw value is $0$ ; if you are finding area, always finish with $\tfrac12|\cdot|$ .

Does the order in which I list the three vertices change the answer?

It can change the sign of the raw determinant, but never the final area, because you take the absolute value at the end. Listing vertices anticlockwise typically gives a positive determinant and listing them clockwise gives a negative one before scaling; either way, $\tfrac12|\cdot|$ lands on the same positive area. For the collinearity test the order does not matter at all, since a zero determinant stays zero under a row swap that merely changes sign.

Can I use this same determinant to check if four points are collinear, or if points in 3-D space lie on a line?

Not directly. The $3\times3$ determinant with rows $(x_i,y_i,1)$ is built specifically for three points in the two-dimensional Cartesian plane. For four points, you would need to test each combination of three of them, or better, check that all the connecting slopes are equal. For points in three dimensions, this exact formula does not apply — collinearity in 3-D is tested instead using the cross product of direction vectors, which is a Vector Algebra topic, not this determinant unit.