Throughout this note we work in .
Why the sign matters
Definition 7.4.1
(Signed area of an ordered pair)
For vectors and in , the signed area of the ordered pair is
The sign is positive when has the same orientation as the standard ordered basis , negative when it has the opposite orientation, and zero when are linearly dependent.
Its absolute value is the area of the parallelogram spanned by and . To see this, expand , where is the angle between and ; hence , which is base times height. Interchanging the vectors reverses the sign. If the determinant is zero, the vectors are dependent and the parallelogram collapses to a line or a point.
Concept lensStructural
Signed area as the alternating bilinear form with value 1 on the standard basis
The formula is forced by three properties of the ordered pair: is linear in each vector separately, , and . The second property gives . Expanding and then leaves only . So signed area is not one formula among several: it is the only function with these properties.
Area scaling under a linear map
A linear map with matrix sends the sides to . Since , the product rule gives
Proposition 7.4.2
(Signed area scales by the determinant)
Let and let be the vertices of a triangle, in this order. Put and . Then the triangle with vertices has signed parallelogram area
so its (ordinary) area is times the original area, and its orientation is the original one if , the opposite one if , and the image is degenerate if .
Proof
(Differences of images are images of differences)
By linearity, and , so the sides of the image triangle at the vertex are and . Their determinant is by the product rule. A triangle has half the area of the parallelogram on the same two sides, in the original and in the image, so the factor passes to triangles unchanged. Taking absolute values gives the factor for ordinary area, and the sign of the product is the sign of times the sign of , which is the orientation statement.
Thus signed area is multiplied by , while ordinary area is multiplied by . A negative factor reverses orientation; zero collapses every planar area. A translation changes neither side vector, so the same rule applies to triangles not based at the origin: with vertices , use sides and divide the parallelogram area by two. Do not take the absolute value until orientation has been determined.
Counterexample mode
Area-preserving does not mean length-preserving
A tempting reading of the proposition is that a matrix with leaves the picture unchanged. Take the shear , so and areas and orientations are preserved. Yet is sent to , and
The unit square, of area , becomes the parallelogram with sides and , whose signed area is , but it is no longer a square. The determinant records only the net area factor; preserving lengths and angles requires the stronger condition .
Worked examples
Worked example 7.4.3
(Area and orientation of an image triangle)
The triangle has vertices , and . Find its area and the area and orientation of its image under .
The ordered sides are and . Their determinant is , so the original orientation is positive and its area is . Now , so the image has signed parallelogram area and triangle area ; its orientation is negative.
For a direct check, and , whose determinant is . Transforming the vertices separately is unnecessary because .
Worked example 7.4.4
(Finding the parameter that triples area)
Let . A triangle with positive orientation has area . For which real does its image have area ? In each case determine the image orientation, and explain what happens at .
The area scaling factor must be , so . The equation gives ; would give , so it has no real solution. At both solutions the determinant is , and the image orientation remains positive. At the determinant is zero: the columns are dependent and the image lies in a line, so every triangle has image area zero. These parameter values do not solve the requested area equation. A negative determinant can reverse orientation for , but here its magnitude is at most , insufficient to triple area.
Worked example 7.4.5
(Vertex order and collinear vertices)
Take , , . The sides are and , with : positive orientation, area . List the same vertices as instead: the sides are and , with determinant . The area is unchanged, and only the orientation (the order in which the vertices are traversed) has flipped.
Now take , , . The sides and give , so the three points are collinear and the "triangle" has area zero. An area formula that returns zero is not an error here; it reports a degenerate triangle, and every invertible linear map sends it to another degenerate one, since .
Common mistakes
Common mistake
(A negative area)
For the matrix with , a triangle of area does not map to a triangle of area . Ordinary area is never negative: it is multiplied by and becomes . The negative sign belongs to the signed area and records only that the orientation has been reversed.
Common mistake
(Forgetting the half)
The number is the area of the parallelogram on the two sides, not of the triangle. The triangle with vertices has area , while the determinant of its sides is . The factor is the same for both, so in a ratio the half cancels, but not in an absolute area.
Quick checks
Quick check
(A linear map of the plane has . A triangle of area is mapped by it. What is the area of the image, and is the orientation preserved?)
Multiply the area by and look at the sign of .
Solution · Answer
The image has area . The determinant is negative, so the orientation is reversed.
Quick check
(Find the signed area of the ordered pair , , and of the pair .)
Use with and .
Solution · Answer
Here , so . Swapping the columns gives . The parallelogram is the same, with area ; only the orientation differs.
Quick check
(Can an invertible linear map of the plane preserve all areas and still reverse orientation? Give a matrix or explain why not.)
Look for a matrix with .
Solution · Answer
Yes. For we have , so areas are multiplied by and orientation is reversed. Indeed is sent to : the map is the reflection in the line .