Volume Factors from Simple Linear Maps
A determinant records more than whether a square matrix is invertible. It gives the signed factor by which the associated linear map changes oriented volume: the absolute value measures the volume change, while the sign records whether orientation is preserved or reversed. Shears and reflections are two useful transformations that make these roles especially clear. A shear changes the position of points by sliding them along a direction, yet its determinant is \(1\). A reflection flips points across a hyperplane, and its determinant is \(-1\).
All matrices and vectors here are real, and vectors in \(\mathbb{R}^n\) are columns. For a matrix \(\mathbf{A}\), write \(T_{\mathbf{A}}(\mathbf{x})=\mathbf{A}\mathbf{x}\) for its associated linear map. By the volume-scaling result for linear maps, a parallelepiped and its image under \(T_{\mathbf{A}}\) have volumes in the ratio \(|\det\mathbf{A}|\). When \(\det\mathbf{A}\ne0\), a positive determinant means orientation is preserved, and a negative determinant means orientation is reversed. In particular, determinant \(1\) means volume and orientation are preserved; determinant \(-1\) means volume magnitude is preserved but orientation is reversed.
A Determinant Formula for Rank-One Changes
Both transformations can be represented as the identity matrix plus a matrix of the form \(\mathbf{a}\mathbf{b}^{\top}\). Such a product has rank at most one: each of its columns is a scalar multiple of \(\mathbf{a}\). The following formula will let us compute its determinant directly.
Proof. The \(j\)th column of \(\mathbf{I}_n+\mathbf{a}\mathbf{b}^{\top}\) is \(\mathbf{e}_j+b_j\mathbf{a}\), where \(b_j\) is the \(j\)th entry of \(\mathbf{b}\). Expand the determinant using linearity in each column. The term using \(\mathbf{e}_j\) in every column is \(\det\mathbf{I}_n=1\). Any term that uses \(\mathbf{a}\) in two or more columns is zero, because those columns are proportional and the determinant is alternating. The terms that use \(\mathbf{a}\) in exactly one column \(j\) contribute \(b_j\det(\mathbf{I}_n\text{ with column }j\text{ replaced by }\mathbf{a})\). Write \(\mathbf{a}=\sum_{i=1}^n a_i\mathbf{e}_i\). In the replaced-column determinant, every term with \(i\ne j\) has two copies of \(\mathbf{e}_i\) among its columns and is zero. The term with \(i=j\) is \(a_j\det\mathbf{I}_n=a_j\). Thus the contribution from column \(j\) is \(b_ja_j\). Adding the unchanged term and all the one-column contributions gives \(1+\sum_{j=1}^n b_ja_j=1+\mathbf{b}^{\top}\mathbf{a}\). \(\square\)
The formula applies to any vectors \(\mathbf{a}\) and \(\mathbf{b}\), including zero vectors. No invertibility assumption is needed. In the applications below, the inner product \(\mathbf{b}^{\top}\mathbf{a}\) determines the determinant immediately.
Shears: Sliding Without Changing Volume
The condition \(\mathbf{v}^{\top}\mathbf{u}=0\) means the direction of the slide lies within the hyperplane \(\{\mathbf{x}:\mathbf{v}^{\top}\mathbf{x}=0\}\). Points in that hyperplane do not move, because their multiplier \(\mathbf{v}^{\top}\mathbf{x}\) is zero. Other points slide by an amount that depends on their value of \(\mathbf{v}^{\top}\mathbf{x}\). The standard coordinate shear \(x_i\mapsto x_i+cx_j\), with all other coordinates unchanged and \(i\ne j\), is obtained by taking \(\mathbf{u}=\mathbf{e}_i\) and \(\mathbf{v}=\mathbf{e}_j\).
Proof. Apply the rank-one determinant formula with \(\mathbf{a}=c\mathbf{u}\) and \(\mathbf{b}=\mathbf{v}\). It gives \(\det\mathbf{S}=1+\mathbf{v}^{\top}(c\mathbf{u})=1+c(\mathbf{v}^{\top}\mathbf{u})=1\). To verify invertibility directly, multiply \(\mathbf{S}\) by \(\mathbf{I}_n-c\mathbf{u}\mathbf{v}^{\top}\). Since \((\mathbf{u}\mathbf{v}^{\top})(\mathbf{u}\mathbf{v}^{\top})=\mathbf{u}(\mathbf{v}^{\top}\mathbf{u})\mathbf{v}^{\top}=\mathbf{O}\), the product is \(\mathbf{I}_n-c\mathbf{u}\mathbf{v}^{\top}+c\mathbf{u}\mathbf{v}^{\top}-c^2\mathbf{O}=\mathbf{I}_n\). The same calculation in the opposite order gives \(\mathbf{I}_n\), so the inverse is \(\mathbf{I}_n-c\mathbf{u}\mathbf{v}^{\top}\). Finally, the volume-scaling result gives volume factor \(|\det\mathbf{S}|=1\), and the positive determinant shows that orientation is preserved. \(\square\)
Worked Example: A Two-Dimensional Shear
Consider the shear that maps \((x_1,x_2)^{\top}\) to \((x_1+3x_2,x_2)^{\top}\). Its matrix is \(\mathbf{S}=\begin{bmatrix}1&3\\0&1\end{bmatrix}\).
The determinant calculation is \(\det\mathbf{S}=1\cdot1-3\cdot0=1\). For a direct check of its action, take \(\mathbf{x}=\begin{bmatrix}2\\-1\end{bmatrix}\). Then \(\mathbf{S}\mathbf{x}=\begin{bmatrix}1&3\\0&1\end{bmatrix}\begin{bmatrix}2\\-1\end{bmatrix}=\begin{bmatrix}2-3\\-1\end{bmatrix}=\begin{bmatrix}-1\\-1\end{bmatrix}\). This agrees with the coordinate rule \((x_1,x_2)\mapsto(x_1+3x_2,x_2)\).
The proposed inverse is \(\begin{bmatrix}1&-3\\0&1\end{bmatrix}\). Multiplication checks it: \(\begin{bmatrix}1&3\\0&1\end{bmatrix}\begin{bmatrix}1&-3\\0&1\end{bmatrix}=\begin{bmatrix}1&0\\0&1\end{bmatrix}\). Thus this map is invertible and, since its determinant is \(1\), it preserves both area and orientation.
Worked Example: A Three-Dimensional Shear
Let \(\mathbf{u}=\begin{bmatrix}1\\0\\1\end{bmatrix}\), \(\mathbf{v}=\begin{bmatrix}0\\1\\0\end{bmatrix}\), and \(c=2\). Since \(\mathbf{v}^{\top}\mathbf{u}=0\), the map is a shear. Its matrix is \(\mathbf{S}=\mathbf{I}_3+2\mathbf{u}\mathbf{v}^{\top}=\begin{bmatrix}1&2&0\\0&1&0\\0&2&1\end{bmatrix}\).
Expanding along the first row gives \(\det\mathbf{S}=1(1\cdot1-0\cdot2)-2(0\cdot1-0\cdot0)+0=1\). For \(\mathbf{x}=\begin{bmatrix}1\\2\\-1\end{bmatrix}\), matrix multiplication gives \(\mathbf{S}\mathbf{x}=\begin{bmatrix}1+4\\2\\4-1\end{bmatrix}=\begin{bmatrix}5\\2\\3\end{bmatrix}\). The coordinate rule gives \(\mathbf{x}+2x_2\mathbf{u}=\begin{bmatrix}1\\2\\-1\end{bmatrix}+4\begin{bmatrix}1\\0\\1\end{bmatrix}=\begin{bmatrix}5\\2\\3\end{bmatrix}\), as required.
The inverse matrix is \(\mathbf{I}_3-2\mathbf{u}\mathbf{v}^{\top}=\begin{bmatrix}1&-2&0\\0&1&0\\0&-2&1\end{bmatrix}\). Multiplying verifies the inverse: \(\begin{bmatrix}1&2&0\\0&1&0\\0&2&1\end{bmatrix}\begin{bmatrix}1&-2&0\\0&1&0\\0&-2&1\end{bmatrix}=\begin{bmatrix}1&0&0\\0&1&0\\0&0&1\end{bmatrix}\). The determinant and volume factor are both \(1\), so this shear preserves three-dimensional volume and orientation.
Reflections: Reversing Orientation
The denominator is nonzero because \(\mathbf{n}\ne\mathbf{0}\). The matrix subtracts twice the component of \(\mathbf{x}\) in the normal direction \(\mathbf{n}\). If \(\mathbf{n}^{\top}\mathbf{x}=0\), then \(\mathbf{H}\mathbf{x}=\mathbf{x}\), so every point in the reflecting hyperplane is fixed. In the normal direction, \(\mathbf{H}\mathbf{n}=\mathbf{n}-2\mathbf{n}=-\mathbf{n}\), so the normal component changes sign.
Proof. Apply the rank-one determinant formula with \(\mathbf{a}=-2\mathbf{n}/(\mathbf{n}^{\top}\mathbf{n})\) and \(\mathbf{b}=\mathbf{n}\). Then \(\det\mathbf{H}=1+\mathbf{n}^{\top}\left(-2\mathbf{n}/(\mathbf{n}^{\top}\mathbf{n})\right)=1-2=-1\). Thus the volume factor is \(|\det\mathbf{H}|=1\), and the negative determinant means orientation is reversed. The map is indeed a reflection: setting \(\mathbf{P}=\mathbf{n}\mathbf{n}^{\top}/(\mathbf{n}^{\top}\mathbf{n})\), direct multiplication gives \(\mathbf{P}^2=\mathbf{P}\), so \(\mathbf{H}^2=(\mathbf{I}_n-2\mathbf{P})^2=\mathbf{I}_n-4\mathbf{P}+4\mathbf{P}^2=\mathbf{I}_n\). Applying the reflection twice therefore returns every vector to itself. \(\square\)
Worked Example: Reflection Across a Plane in R³
Take \(\mathbf{n}=\begin{bmatrix}1\\1\\0\end{bmatrix}\), so the reflecting plane is \(x_1+x_2=0\). Since \(\mathbf{n}^{\top}\mathbf{n}=2\), the reflection matrix is \(\mathbf{H}=\mathbf{I}_3-\mathbf{n}\mathbf{n}^{\top}=\begin{bmatrix}0&-1&0\\-1&0&0\\0&0&1\end{bmatrix}\).
Its determinant is \(\det\mathbf{H}=0-(-1)\bigl((-1)\cdot1-0\cdot0\bigr)+0=-1\). For \(\mathbf{x}=\begin{bmatrix}2\\1\\3\end{bmatrix}\), the image is \(\mathbf{H}\mathbf{x}=\begin{bmatrix}-1\\-2\\3\end{bmatrix}\). The sum of the first two coordinates changes from \(3\) to \(-3\), while the third coordinate remains unchanged. A vector in the plane, such as \(\begin{bmatrix}1\\-1\\0\end{bmatrix}\), is fixed: \(\mathbf{H}\begin{bmatrix}1\\-1\\0\end{bmatrix}=\begin{bmatrix}1\\-1\\0\end{bmatrix}\).
Multiplication also confirms that applying the reflection twice returns the original vector: \(\mathbf{H}^2=\begin{bmatrix}0&-1&0\\-1&0&0\\0&0&1\end{bmatrix}\begin{bmatrix}0&-1&0\\-1&0&0\\0&0&1\end{bmatrix}=\begin{bmatrix}1&0&0\\0&1&0\\0&0&1\end{bmatrix}\). The negative determinant indicates reversed orientation, not negative physical volume. The volume magnitude is unchanged because \(|-1|=1\).
Distinguishing Volume from Orientation
These examples illustrate why the determinant must be interpreted using both its sign and its absolute value. The volume of a region is nonnegative, so the scaling factor for volume is \(|\det\mathbf{A}|\). The sign of \(\det\mathbf{A}\) instead records whether the ordered coordinate directions keep their orientation. A reflection does not create a negative volume: it reverses orientation while leaving the volume magnitude unchanged.
A shear provides a different geometric picture. In the coordinate shear \((x_1,x_2)\mapsto(x_1+cx_2,x_2)\), points with \(x_2=0\) stay fixed, and other horizontal layers slide. The shape of a region may change, but the signed area factor remains \(1\). In higher dimensions, the same principle holds for a shear along any direction lying in the fixed hyperplane. The determinant calculation makes the volume conclusion exact, without relying on a drawing or on a special choice of parallelepiped.
Common Pitfall: A Negative Determinant Is Not Negative Volume
A common mistake is to interpret \(\det\mathbf{H}=-1\) as saying that a reflection changes a region’s volume to \(-1\) times its original volume. The determinant is a signed volume factor, but ordinary volume uses its magnitude. For a region of volume \(V\), reflection gives volume \(|-1|V=V\), while the negative sign records reversal of orientation.
Another tempting error is to read an off-diagonal shear entry as a volume change. For example, one might wrongly claim that \(\begin{bmatrix}1&3\\0&1\end{bmatrix}\) has determinant \(1-3=-2\), treating the off-diagonal \(3\) as if it were a diagonal contribution. The correct two-by-two formula gives \(\det\begin{bmatrix}1&3\\0&1\end{bmatrix}=1\cdot1-3\cdot0=1\). The shear’s nonzero off-diagonal entry changes the map’s action, but it does not by itself change its volume factor.
| Transformation | Determinant | Volume magnitude | Orientation |
|---|---|---|---|
| Shear with \(\mathbf{v}^{\top}\mathbf{u}=0\) | \(1\) | Preserved | Preserved |
| Reflection across a hyperplane | \(-1\) | Preserved | Reversed |
Check Your Understanding
Use the defining matrices and the determinant’s signed volume interpretation to answer each question.
- Write the matrix for \((x_1,x_2)\mapsto(x_1-4x_2,x_2)\), and calculate its determinant.
- For \(\mathbf{u}=\begin{bmatrix}1\\0\\1\end{bmatrix}\) and \(\mathbf{v}=\begin{bmatrix}1\\0\\-1\end{bmatrix}\), check whether \(\mathbf{v}^{\top}\mathbf{u}=0\). If so, what is the determinant of \(\mathbf{I}_3+5\mathbf{u}\mathbf{v}^{\top}\)?
- Find the reflection matrix across the line \(x_1=0\) in \(\mathbb{R}^2\), and state its determinant and volume factor.
- State the rank-one determinant formula for \(\det(\mathbf{I}_n+\mathbf{a}\mathbf{b}^{\top})\).
- Explain why a determinant of \(-1\) means orientation is reversed but volume is preserved.
