Determinants

⚡ Key Takeaways

A determinant is a single number computed from a square matrix that tells you whether the matrix is invertible (nonzero) or singular (zero), and measures how the linear transformation scales area or volume. Key result: a system $A\mathbf{x}=\mathbf{b}$ has a unique solution if and only if $\det(A)\neq 0$ . Applies across linear algebra, geometry, calculus (Jacobians), and physics. Introduced by Leibniz (1693) and Cramer (1750); named by Cauchy (1812).

💡 The Core Insight

The determinant is the signed scaling factor for area (2D) or volume (3D) under the linear transformation defined by the matrix. If you take a unit square and apply the matrix, the resulting parallelogram has area $|\det(A)|$ . When $\det(A)=0$ , the transformation collapses space into a lower dimension (a line or a point), destroying information — which is exactly why no inverse can exist.

DefinitionDeterminant of a 2×2 Matrix

For

1

\quad A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}, \qquad \det(A) = ad - bc.

Derivation hint: This is the signed area of the parallelogram spanned by the column vectors $(a,c)$ and $(b,d)$ ; the cross-product-style formula gives the oriented area. Here $a,b,c,d \in \mathbb{R}$ are the matrix entries. The notation $\det(A)$ , $|A|$ , and $\begin{vmatrix}a&b\\c&d\end{vmatrix}$ all mean the same number.

DefinitionDeterminant of a 3×3 Matrix (Sarrus / Cofactor)

For a $3\times 3$ matrix

2

\quad A=\begin{pmatrix} a_{11}&a_{12}&a_{13}\\ a_{21}&a_{22}&a_{23}\\ a_{31}&a_{32}&a_{33}\end{pmatrix},

3

\quad \det(A)=a_{11}\begin{vmatrix}a_{22}&a_{23}\\a_{32}&a_{33}\end{vmatrix}-a_{12}\begin{vmatrix}a_{21}&a_{23}\\a_{31}&a_{33}\end{vmatrix}+a_{13}\begin{vmatrix}a_{21}&a_{22}\\a_{31}&a_{32}\end{vmatrix}.

Derivation hint: Expand along the first row; each entry multiplies the determinant of the $2\times 2$ minor obtained by deleting its row and column, with alternating signs $+,-,+$ . Each $a_{ij}$ denotes the entry in row $i$ , column $j$ . The $2\times 2$ determinants are called minors, and the signed minors are cofactors.

DefinitionGeneral n×n Determinant — Cofactor (Laplace) Expansion

For an $n\times n$ matrix, expanding along row $i$ :

4

\quad \det(A)=\sum_{j=1}^{n}(-1)^{i+j}\,a_{ij}\,M_{ij},

Derivation hint: $M_{ij}$ is the minor — the determinant of the $(n-1)\times(n-1)$ matrix left after deleting row $i$ and column $j$ . The factor $(-1)^{i+j}$ produces the checkerboard sign pattern. where $M_{ij}$ is the minor of entry $a_{ij}$ and $C_{ij}=(-1)^{i+j}M_{ij}$ is its cofactor. The expansion gives the same value along any row or any column.

DefinitionLeibniz Formula (Permutation Definition)

The fully expanded definition is

5

\quad \det(A)=\sum_{\sigma\in S_n}\operatorname{sgn}(\sigma)\prod_{i=1}^{n}a_{i,\sigma(i)},

Derivation hint: Sum over all $n!$ permutations $\sigma$ of $\{1,\dots,n\}$ ; each term is a product picking exactly one entry from each row and each column, weighted by the sign of the permutation ( $+1$ for even, $-1$ for odd). $S_n$ is the set of all permutations of $n$ symbols, and $\operatorname{sgn}(\sigma)=\pm 1$ .

TheoremInvertibility Criterion

A square matrix $A$ is invertible if and only if $\det(A)\neq 0$ . When invertible,

6

\quad \det(A^{-1})=\frac{1}{\det(A)}.

Derivation hint: Apply the multiplicativity rule $\det(AB)=\det(A)\det(B)$ to $AA^{-1}=I$ , using $\det(I)=1$ .

TheoremMultiplicativity

For any two $n\times n$ matrices,

7

\quad \det(AB)=\det(A)\,\det(B).

Derivation hint: Both sides measure how the composite transformation scales volume; composing two transformations multiplies their scaling factors. A rigorous proof uses elementary matrices.

ProofThe 2×2 Determinant as Signed Area

We show $\det(A)=ad-bc$ equals the signed area of the parallelogram spanned by $\mathbf{u}=(a,c)$ and $\mathbf{v}=(b,d)$ . 1. The area of a parallelogram with sides $\mathbf{u},\mathbf{v}$ is $|\mathbf{u}||\mathbf{v}|\sin\theta$ , where $\theta$ is the angle between them. 2. This equals the magnitude of the 2D cross product, $u_x v_y - u_y v_x = a\cdot d - c\cdot b = ad-bc$ . 3. The sign records orientation: positive if $\mathbf{v}$ lies counterclockwise from $\mathbf{u}$ , negative if clockwise. 4. Hence $\det(A)=ad-bc$ is exactly the signed area. If the two columns are parallel (linearly dependent), the parallelogram degenerates to a segment of zero area, giving $\det(A)=0$ . $\blacksquare$

∎

ProofWhy det = 0 Implies Non-Invertibility

1. Suppose $\det(A)=0$ . Geometrically, the parallelepiped formed by the columns of $A$ has zero volume. 2. Zero volume means the column vectors are linearly dependent — they lie in a lower-dimensional subspace. 3. Therefore the transformation $\mathbf{x}\mapsto A\mathbf{x}$ maps all of space into that subspace, collapsing dimension. 4. A collapse cannot be undone: distinct inputs map to the same output, so no inverse function exists. 5. Conversely, if $\det(A)\neq 0$ , no collapse occurs, the columns span all of space, and $A^{-1}$ exists. $\blacksquare$

∎

Figure

Determinant as the signed area of the parallelogram spanned by a matrix's columns

mathematician · Augustin-Louis Cauchy | 1789–1857

Although the idea of determinants predates him by over a century, Cauchy gave the subject its modern name and systematic foundation. In an 1812 memoir he coined the word "déterminant," proved the multiplication theorem $\det(AB)=\det(A)\det(B)$ in general, and organized the scattered earlier results into a coherent theory. His treatment unified the work of Leibniz, Cramer, Vandermonde, and Laplace, transforming determinants from a computational trick for solving equations into a central object of linear algebra.

🧠 Intuition

Think of a matrix as a machine that stretches, squishes, rotates, and shears space. The determinant is the single number that reports how much that machine changes area or volume.

$\det = 2$ : areas double.
$\det = 1$ : areas are preserved (a rotation or shear).
$\det = 0.5$ : areas shrink to half.
$\det = 0$ : space is flattened onto a line or point — irreversible.
$\det = -1$ : areas preserved, but space is flipped (like a mirror reflection).

The negative sign is the part students miss: it has nothing to do with size and everything to do with orientation. A clockwise-labelled triangle becomes counterclockwise. This is why $\det$ is called the signed volume scaling factor.

ExampleA 3×3 Determinant by Cofactor Expansion

Given:

\begin{pmatrix} 2 & -1 & 3 \\ 0 & 4 & 5 \\ 1 & 0 & 6 \end{pmatrix}

Steps: Expand along the first row using equation [3]. $\det = 2\begin{vmatrix}4&5\\0&6\end{vmatrix} - (-1)\begin{vmatrix}0&5\\1&6\end{vmatrix} + 3\begin{vmatrix}0&4\\1&0\end{vmatrix}$

$\begin{vmatrix}4&5\\0&6\end{vmatrix}=4\cdot6-5\cdot0=24$
$\begin{vmatrix}0&5\\1&6\end{vmatrix}=0\cdot6-5\cdot1=-5$
$\begin{vmatrix}0&4\\1&0\end{vmatrix}=0\cdot0-4\cdot1=-4$

So $\det = 2(24) + 1(-5) + 3(-4) = 48 - 5 - 12 = 31$ .

Answer:

\boxed{\det(A) = 31 \text{ (dimensionless)}}

Since $31\neq 0$ , the matrix is invertible.

ExampleSolving a System with Cramer's Rule

Given: Solve

\begin{cases} 3x + 2y = 7 \\ x - y = 1 \end{cases}

Steps: The coefficient matrix is $A=\begin{pmatrix}3&2\\1&-1\end{pmatrix}$ , with $\det(A)=3(-1)-2(1)=-5$ . Replace column 1 with the constants $(7,1)$ :

\det(A_x)=\begin{vmatrix}7&2\\1&-1\end{vmatrix}=7(-1)-2(1)=-9.

Replace column 2 with the constants:

\det(A_y)=\begin{vmatrix}3&7\\1&1\end{vmatrix}=3(1)-7(1)=-4.

Then $x=\dfrac{\det(A_x)}{\det(A)}=\dfrac{-9}{-5}=\dfrac{9}{5}$ and $y=\dfrac{\det(A_y)}{\det(A)}=\dfrac{-4}{-5}=\dfrac{4}{5}$ .

Answer:

\boxed{x = \tfrac{9}{5} = 1.8, \quad y = \tfrac{4}{5} = 0.8}

Check: $3(1.8)+2(0.8)=5.4+1.6=7$ ✓.

⚠️ Common mistake

The determinant is NOT additive. In general,

\det(A+B)\neq \det(A)+\det(B).

It is multiplicative ( $\det(AB)=\det(A)\det(B)$ ) but never additive. Also note: $\det(kA)=k^n\det(A)$ for an $n\times n$ matrix — scaling every entry by $k$ multiplies the determinant by $k^n$ , not by $k$ , because all $n$ dimensions are scaled.

Counterexample

Let $A=\begin{pmatrix}1&0\\0&1\end{pmatrix}$ and $B=\begin{pmatrix}-1&0\\0&-1\end{pmatrix}$ . Then $\det(A)=1$ and $\det(B)=1$ , so $\det(A)+\det(B)=2$ . But $A+B=\begin{pmatrix}0&0\\0&0\end{pmatrix}$ , giving $\det(A+B)=0\neq 2$ . This kills any temptation to "distribute" the determinant over addition.

⚠️ Common mistake

Determinants are only defined for square matrices. A $2\times 3$ or $3\times 2$ matrix has no determinant. Likewise, a determinant being nonzero does not tell you the actual solution — it only guarantees a unique solution exists. You still need Cramer's Rule, row reduction, or the inverse to find it.

connection

The determinant ties together nearly every thread of linear algebra. It detects whether a matrix is invertible (see Matrix Inverse) and whether a System of Linear Equations has a unique solution. It appears in the characteristic equation $\det(A-\lambda I)=0$ that defines Eigenvalues and Eigenvectors. In geometry it is the signed area/volume, linking to the Cross Product and Vectors. In multivariable Calculus the Jacobian determinant governs how integrals transform under a change of variables. It is also the foundation of Cramer's Rule and the cofactor formula for the Matrix Inverse.

🔬 Where this appears

⏳ History

- 1683 — Seki Takakazu in Japan independently formulates determinant-like expressions for solving simultaneous equations.

- 1693 — Gottfried Wilhelm Leibniz describes determinants in a letter to L'Hôpital, using an early index notation.

- 1750 — Gabriel Cramer publishes Cramer's Rule, giving explicit determinant formulas for solving linear systems.

- 1772 — Pierre-Simon Laplace develops the cofactor (Laplace) expansion, generalizing computation to any size.

- 1773 — Joseph-Louis Lagrange connects determinants to volumes of tetrahedra and parallelepipeds.

- 1812 — Augustin-Louis Cauchy coins the term déterminant, proves multiplicativity, and systematizes the theory.

- 1841 — James Joseph Sylvester and Arthur Cayley advance matrix and determinant theory; Carl Jacobi popularizes the Jacobian determinant.

- 1858 — Cayley's Memoir on the Theory of Matrices formally separates the matrix as an object from its determinant.

open-problem

For large matrices, computing the determinant via the Leibniz formula [5] requires summing $n!$ terms — astronomically slow ( $20! \approx 2.4\times 10^{18}$ ). Practical computation uses LU decomposition via Gaussian elimination ( $\sim n^3$ operations), but this introduces numerical instability: for ill-conditioned matrices, rounding errors can make a computed determinant wildly inaccurate, sometimes returning a tiny nonzero value for a matrix that is effectively singular. This is why numerical analysts often prefer the condition number or singular values over the raw determinant to judge invertibility in practice. The interplay between exact symbolic determinants and numerically stable approximations remains a core concern in computational linear algebra.

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also