Determinants in Geometry

Three villages — Rampur, Sitapur, and Lakshmipur — sit at coordinates (1, 2), (4, 6), and (7, 4) on a survey map. A government engineer needs the area of the triangular region connecting them, to estimate the cost of laying a road network. She could use the base-height formula from school, but which side is the base? And how does she compute the perpendicular height to that base from the third vertex?

There is a far cleaner method. Write the coordinates into a 3 \times 3 determinant:

Expand this determinant (along the third column, say), and you get a single number. Half its absolute value is the area. No base, no height, no perpendicular — just coordinates plugged into a formula.

That is what this article is about: the surprising power of determinants as a tool for coordinate geometry.

Area of a triangle using determinants

Given three points P_1 = (x_1, y_1), P_2 = (x_2, y_2), P_3 = (x_3, y_3), the area of the triangle they form is:

Where the formula comes from

The derivation starts from the coordinate geometry formula for the area of a triangle with vertices (x_1, y_1), (x_2, y_2), (x_3, y_3):

This is a standard result you may have seen derived using the "shoelace" method — splitting the triangle into vertical strips.

Now expand the determinant \begin{vmatrix} x_1 & y_1 & 1 \\ x_2 & y_2 & 1 \\ x_3 & y_3 & 1 \end{vmatrix} along the third column:

Why: expanding along the third column is the natural choice because every entry in that column is 1, which simplifies the cofactors. The result is exactly the coordinate geometry formula for twice the signed area.

The determinant gives the signed area — positive if the vertices go anticlockwise, negative if clockwise. The absolute value strips the sign, leaving the area.

Area of a quadrilateral

For a quadrilateral with vertices P_1, P_2, P_3, P_4 (in order), split it into two triangles along a diagonal — say P_1 P_3. The area is the sum of the two triangle areas:

For a convex quadrilateral, this always works regardless of which diagonal you choose. The determinant approach scales naturally to any polygon by triangulating from a single vertex.

Condition for collinearity

Three points are collinear (they lie on the same straight line) if and only if the triangle they form has zero area. From the area formula, this means:

Why: zero area means the three points do not form a triangle — they are squeezed onto a single line. The determinant captures this exactly: it vanishes when the three rows are linearly dependent, which is the algebraic way of saying the points are collinear.

This is more than just a restatement of the area formula. The determinant form is cleaner — no fractions, no absolute value. The condition is simply: determinant equals zero. This makes it easy to use in proofs and in problems where you are checking a geometric property.

For example, are the points (1, 5), (3, 9), (5, 13) collinear?

Yes, the determinant is zero, so the points are collinear. You can verify: they all lie on the line y = 2x + 3.

Equation of a line through two points

Here is where the determinant becomes a tool for writing equations, not just checking conditions.

Suppose you want the equation of the line through (x_1, y_1) and (x_2, y_2). An arbitrary point (x, y) lies on this line if and only if it is collinear with the two given points. The collinearity condition gives the equation directly:

Why this works

The determinant is zero precisely when the three points (x, y), (x_1, y_1), (x_2, y_2) are collinear — and a point (x, y) is collinear with two fixed points exactly when it lies on the line through them.

Expand the determinant:

This is a linear equation in x and y — exactly the equation of a straight line. You can rearrange it into slope-intercept form or any other standard form.

For example, the line through (2, 3) and (5, 9):

Dividing by 3: -2x + y + 1 = 0, or y = 2x - 1.

Condition for concurrency

Three lines are concurrent if they all pass through a single point. Suppose the three lines are:

Why the determinant condition works

If the three lines are concurrent, they meet at some point (p, q). Then:

This is a system of three linear equations in p and q — but it's actually saying that the vector (p, q, 1) is a non-trivial solution to:

A homogeneous system has a non-trivial solution if and only if the determinant of the coefficient matrix is zero.

Why: three lines in a plane can be concurrent (all meet at one point), form a triangle (each pair meets at a different point), or include parallel lines. The determinant being zero picks out exactly the first case — the rows are linearly dependent, meaning one line's equation is a combination of the other two, which forces all three to share a common point.

A quick check

Are the lines x + y - 1 = 0, 2x - y - 5 = 0, and x - 2y + 2 = 0 concurrent?

Expand along the first row:

The determinant is -18 \neq 0, so the lines are not concurrent — they form a triangle instead. Each pair of lines meets at a different point, and no single point lies on all three.

Now try three lines that are concurrent: x + y = 3, 2x - y = 0, and x - 2y = -3. In standard form these are x + y - 3 = 0, 2x - y + 0 = 0, and x - 2y + 3 = 0.

Expand along the first row:

The determinant is zero — the lines are concurrent. You can verify: solving the first two equations gives x = 1, y = 2, and plugging into the third: 1 - 4 + 3 = 0. All three lines pass through (1, 2).

The concurrency condition is especially useful in triangle geometry. The three medians of a triangle are concurrent (they meet at the centroid). The three altitudes are concurrent (they meet at the orthocentre). The three perpendicular bisectors are concurrent (they meet at the circumcentre). In each case, writing the three lines in general form and checking the determinant condition gives a proof of concurrency — often more compact than the geometric argument.

Equation of a circle through three points

The determinant trick generalises beyond lines. Three non-collinear points determine a unique circle. If the three points are (x_1, y_1), (x_2, y_2), (x_3, y_3), the equation of the circle through them can be written as:

This is a 4 \times 4 determinant in which the first row has the variables x and y (and x^2 + y^2). When you expand it, you get a linear expression in x^2 + y^2, x, and y — which is exactly the general equation of a circle: x^2 + y^2 + Dx + Ey + F = 0.

Why: the determinant is zero when the point (x, y) lies on the same circle as the three given points. Each row encodes one point's relationship to the general circle equation, and the determinant vanishing means all four points satisfy the same circle equation.

A concrete circle example

Find the equation of the circle through (0, 0), (2, 0), and (0, 4).

Expand along the second row. Since three entries in row 2 are zero, only the (2, 4) cofactor survives (with the appropriate sign):

Expand this 3 \times 3 determinant along the third column:

Working through the cofactors carefully:

Setting the whole expression (with the leading -1) to zero:

Dividing by 8: x^2 + y^2 - 2x - 4y = 0, or (x - 1)^2 + (y - 2)^2 = 5. The circle has centre (1, 2) and radius \sqrt{5}.

You can verify: the origin (0, 0) gives 1 + 4 = 5; the point (2, 0) gives 1 + 4 = 5; the point (0, 4) gives 1 + 4 = 5. All three points lie on the circle, as required.

The same approach works for conics. Five points determine a unique conic (ellipse, parabola, or hyperbola), and a 6 \times 6 determinant using x^2, xy, y^2, x, y, 1 as the first row produces the conic's equation.

Worked examples

Common confusions

• "The area can be negative." The determinant can be negative — that tells you the orientation (clockwise vs anticlockwise). The area is always non-negative, which is why you take the absolute value and divide by 2. If you forget the absolute value, you will sometimes get a negative area, which is geometrically meaningless.

• "The 1s in the third column are part of the coordinates." They are not. The 1 is an augmentation — it converts the coordinates into homogeneous form, which is what makes the determinant formula work. The actual coordinates are only the first two columns.

• "Concurrency requires three lines to be identical." No — concurrency means three distinct lines all passing through the same point. The determinant being zero does not mean the lines are the same; it means they share a common intersection. (If two of the lines are actually the same, the determinant is still zero, but for a trivial reason — two identical rows.)

• "The circle formula needs a 3 \times 3 determinant." The line through two points uses a 3 \times 3 determinant with three points. The circle through three points needs a 4 \times 4 determinant — one extra row and column to accommodate the x^2 + y^2 term. The pattern is: n conditions require an (n+1) \times (n+1) determinant.

Going deeper

Signed area and orientation

The determinant \begin{vmatrix} x_1 & y_1 & 1 \\ x_2 & y_2 & 1 \\ x_3 & y_3 & 1 \end{vmatrix} gives twice the signed area of the triangle. The sign encodes the orientation of the vertices:

• Positive: the vertices go anticlockwise (the standard mathematical convention).
• Negative: the vertices go clockwise.
• Zero: the vertices are collinear (the "triangle" degenerates to a line segment).

The signed area is more fundamental than the unsigned area. It is an invariant under translation (shifting all points by the same vector does not change the determinant) and changes sign under reflection. These properties make it central to computational geometry, where algorithms often need to know not just the size of a region but also whether a sequence of points turns left or right.

The general conic through five points

A conic (ellipse, parabola, or hyperbola) has the general equation:

This has six coefficients, but only five ratios matter (you can divide through by any non-zero coefficient). So five points determine a unique conic, and the equation can be written as:

Expanding this 6 \times 6 determinant along the first row gives a second-degree equation in x and y — the equation of the conic. The circle is a special case where the xy coefficient is zero and the x^2 and y^2 coefficients are equal.

The projective viewpoint

The column of 1s in the area determinant is not arbitrary. It comes from homogeneous coordinates: the point (x, y) in the plane is represented as (x, y, 1) in projective space. The determinant \begin{vmatrix} x_1 & y_1 & 1 \\ x_2 & y_2 & 1 \\ x_3 & y_3 & 1 \end{vmatrix} is the determinant of a matrix whose rows are the homogeneous coordinates of three points.

In projective geometry, this determinant vanishes when the three projective points are "dependent" — lying on a single projective line. This is the projective version of collinearity, and it works even for points at infinity (where the third coordinate is 0 instead of 1). The column of 1s is not padding; it is the mathematics of projection.

Cross-ratio and determinants

The cross-ratio of four collinear points, a fundamental invariant in projective geometry, can also be expressed using determinants. If P_1, P_2, P_3, P_4 are four points on a line with homogeneous coordinates as column vectors, then the cross-ratio is:

where [P_i P_j] denotes the 2 \times 2 determinant formed by the homogeneous coordinates. The cross-ratio is the unique quantity preserved by all projective transformations — it is the projective geometer's version of distance.

Brahmagupta's formula for the area of a cyclic quadrilateral and Bhaskara II's work on cyclic polygons both relate, at their core, to determinantal expressions of the kind developed in this article. The coordinates change, but the underlying algebraic structure — rows of a matrix encoding points, and the determinant encoding their relationship — remains the same.

Where this leads next

Determinants in geometry connect to both linear algebra and coordinate geometry. The natural next articles:

• Properties of Determinants — the row and column operations that simplify the computations in this article.
• Product of Determinants — the rule \det(AB) = \det(A)\det(B), which underpins transformations of geometric figures.
• Straight Lines — a full treatment of lines in the coordinate plane, with multiple forms for the equation.
• Special Determinants — Vandermonde, circulant, and other structured determinants with closed-form answers.
• Conic Sections — circles, ellipses, parabolas, and hyperbolas in coordinate geometry.