Determinants as Signed Volume Scale Factors in R3

From a Determinant to a Three-Dimensional Shape

Three vectors in \(\mathbb{R}^3\) can be viewed in two ways at once. As the columns of a matrix, they determine a determinant; as the three edge vectors from one corner, they determine a parallelepiped. The connection between these views is geometric: the absolute value of the determinant is the parallelepiped's volume. The sign contains additional information, recording whether the ordered edges have the standard orientation or the opposite one.

Throughout, vectors are columns in \(\mathbb{R}^3\). Let \(\mathbf{a},\mathbf{b},\mathbf{c}\in\mathbb{R}^3\). The parallelepiped they span consists of points \(s\mathbf{a}+t\mathbf{b}+u\mathbf{c}\), where \(0\le s,t,u\le1\). Its ordinary volume is never negative. Its signed volume, by contrast, is a number whose magnitude is the ordinary volume and whose sign tracks orientation.

To justify this interpretation, we will relate the determinant to the base area and the height above that base. The cross product provides a convenient way to express both. For \(\mathbf{a}=(a_1,a_2,a_3)\) and \(\mathbf{b}=(b_1,b_2,b_3)\), define

The Scalar Triple Product

The dot product of the cross product with the third vector is called the scalar triple product. Its formula is particularly useful here because it involves both the determinant and the component of \(\mathbf{c}\) in the direction perpendicular to the base.

Proof. Write \(\mathbf{a}=(a_1,a_2,a_3)\), \(\mathbf{b}=(b_1,b_2,b_3)\), and \(\mathbf{c}=(c_1,c_2,c_3)\). By the definition of the cross product,

Expanding the determinant of the matrix with these three columns gives the same terms:

The expressions are equal, proving the identity. \(\square\)

We also need the geometric meaning of the cross product's norm. Its algebraic identity is

Indeed, both sides equal \(a_1^2b_2^2+a_1^2b_3^2+a_2^2b_1^2+a_2^2b_3^2+a_3^2b_1^2+a_3^2b_2^2-2a_1a_2b_1b_2-2a_1a_3b_1b_3-2a_2a_3b_2b_3\), as follows by expanding the definitions.

When \(\mathbf{a}\ne\mathbf{0}\), decompose \(\mathbf{b}\) into a component parallel to \(\mathbf{a}\) and a perpendicular component \(\mathbf{b}_{\perp}\). Since \(\mathbf{a}\cdot\mathbf{b}_{\perp}=0\), the identity gives \(\|\mathbf{a}\times\mathbf{b}\|^2=\|\mathbf{a}\|^2\|\mathbf{b}_{\perp}\|^2\), hence \(\|\mathbf{a}\times\mathbf{b}\|=\|\mathbf{a}\|\|\mathbf{b}_{\perp}\|\). This is base times perpendicular height, the area of the parallelogram. If \(\mathbf{a}=\mathbf{0}\), both that area and the cross-product norm are zero as well.

The Signed Volume Formula

Proof. First suppose \(\mathbf{a}\times\mathbf{b}\ne\mathbf{0}\), so \(\mathbf{a}\) and \(\mathbf{b}\) form a nondegenerate base. The unit normal pointing in the direction of \(\mathbf{a}\times\mathbf{b}\) is \(\mathbf{n}=(\mathbf{a}\times\mathbf{b})/\|\mathbf{a}\times\mathbf{b}\|\). The signed component of \(\mathbf{c}\) in this normal direction is \(\mathbf{n}\cdot\mathbf{c}\), so the height of the parallelepiped is \(h=|\mathbf{n}\cdot\mathbf{c}|\). Its base area is \(\|\mathbf{a}\times\mathbf{b}\|\). Therefore

If \(\mathbf{a}\times\mathbf{b}=\mathbf{0}\), the base has zero area and the parallelepiped has zero volume. The scalar triple product is also zero, since \((\mathbf{a}\times\mathbf{b})\cdot\mathbf{c}=0\); by the proposition, the determinant is zero. Thus the volume formula holds in this case too. In the nondegenerate-base case, \(\mathbf{n}\cdot\mathbf{c}\) is positive when \(\mathbf{c}\) points partly in the normal direction and negative when it points partly in the opposite direction. The determinant has the same sign, because it equals \(\|\mathbf{a}\times\mathbf{b}\|(\mathbf{n}\cdot\mathbf{c})\), with a positive factor \(\|\mathbf{a}\times\mathbf{b}\|\). This proves the orientation statement. \(\square\)

In the standard coordinate system, \((\mathbf{e}_1,\mathbf{e}_2,\mathbf{e}_3)\) has positive orientation: \(\mathbf{e}_1\times\mathbf{e}_2=\mathbf{e}_3\), so its scalar triple product with \(\mathbf{e}_3\) is \(1\). A triple with positive determinant has that same orientation; one with negative determinant has the opposite orientation. If the determinant is zero, the height above the base is zero or the base itself has zero area, and there is no nonzero orientation to record.

Worked Examples

What the Sign Does—and Does Not—Mean

A negative determinant does not describe a physically negative amount of space. Ordinary volume is always nonnegative; the absolute value of the determinant gives it. The sign is useful because it keeps track of the order and orientation of the three edge vectors. For example, exchanging two columns reverses the sign of a determinant, while the same geometric parallelepiped still has the same ordinary volume. The sign therefore records ordered orientation, not a change in size.

A zero determinant has a geometric interpretation as well. If the base vectors are parallel or one is zero, the base area is zero. If the base is nondegenerate but the third vector lies in its plane, the height is zero. In either case, the parallelepiped is collapsed into a lower-dimensional shape and has zero volume. The scalar triple product expresses both possibilities without requiring a separate volume calculation.

Common mistake. For the columns in the negative-orientation example, it is wrong to report “the volume is \(-2\).” The determinant is \(-2\), but ordinary volume is its absolute value, \(2\). The corrected interpretation keeps both pieces of information: signed volume \(-2\), ordinary volume \(2\), and negative orientation.

Another useful check is to calculate the base area and height separately when they are simple. If their product disagrees with the absolute value of the determinant, revisit the column order, the determinant expansion, or the perpendicular height. The height is the component normal to the base, not necessarily the length \(\|\mathbf{c}\|\).