determinant as area scaling
When you learned determinants in school, you probably just memorized ad minus bc. Nobody likely told you what it means, yet the determinant is a remarkably intuitive number: it is how many times a transformation scales area, and nothing more. If det is 3, every shape's area triples; if det is 0, the plane has been squashed flat. Once that sentence sinks in, the ad minus bc formula and the reason an inverse sometimes fails to exist both come out of the same picture.
To find how many times area changes, pick one reference shape and see how much it grows. The cleanest reference is the unit square spanned by î and ĵ, whose area is exactly 1. Applying a matrix moves î and ĵ, and as the previous lesson, "Matrices Are Transformations," showed, the two columns of the matrix are where î and ĵ go. So the square is squeezed into a parallelogram. Move the four sliders and the gold parallelogram changes shape, while the determinant and its absolute value, the area, are shown below. Because the original area was 1, the new area is exactly the scale factor, and that is the determinant. When the determinant goes negative the parallelogram turns red; what that means comes shortly. Compute the area of this parallelogram directly from coordinates and you get exactly ad minus bc. The formula you memorized in school was really a way to compute this area.
You might ask whether this works only for the unit square. It does not. Place any shape on the plane and its area changes by the same ratio. A circle or a lumpy blob can be thought of as many tiny squares; each piece grows by the same factor, so the whole does too. The figure holds a green triangle and a blue quadrilateral together. Change the matrix with the sliders and the two shapes deform in different ways, yet the area ratios shown below always agree with each other and with the absolute value of the determinant. So the determinant is not a property of any one shape; it is a single area scale factor applied uniformly to all of space.
Area should be positive, yet a determinant is sometimes negative. What could a negative area mean? It means space has been flipped. Turn a sheet of paper over and the writing reverses as in a mirror; a transformation can flip the plane in the same way. A turn that went counterclockwise from î to ĵ becomes clockwise after the flip. In the figure, the arc and its arrowhead show the direction of the turn from î to ĵ. Push the slider to the left: the moment the determinant goes negative, the arc turns the other way and turns red, and the label changes to "flipped." The negative sign records that the direction has reversed. A single determinant carries both a size, the area scale factor, and a direction, whether space was flipped.
Now for the most important case. The slider in this figure controls the determinant directly. Ease it toward 0 and the parallelogram grows flatter and flatter until it collapses onto a single line. The area has become 0, and two dimensions have been crushed into 1. At that moment a "no inverse" warning appears below. The key point is that once space is flattened, there is no way back. Different points have been squashed onto the same spot, so from the points on the line alone you cannot tell where in the original plane they came from. "A matrix with determinant 0 has no inverse" is really this picture. What has been crushed cannot be unfolded. The story continues properly in the next lesson, on systems of equations and inverses.
What happens to area when you apply two transformations in a row? No need to overthink it. Apply a transformation that doubles area, then one that triples it, and the final area is 2 times 3, or 6 times the original. So the determinant of a product of matrices equals the product of their determinants: det(AB) = det(A) × det(B). People usually memorize this as a formula, but once you remember that scale factors multiply, it is obvious. Use the arrow buttons to apply the first matrix and then the second to the unit square: the area readout below multiplies at each step, and the last line confirms that the product of the two determinants equals the determinant of the combined transformation.