What is Cramer's rule?

Cramer's rule solves Ax = b by expressing each solution component as xᵢ = det(Aᵢ)/det(A), where Aᵢ replaces column i with b. It requires n+1 determinants, so it's slower than elimination for large systems but useful for symbolic work.

How do you find the inverse using the adjugate?

A⁻¹ = adj(A)/det(A), where adj(A) is the transpose of the cofactor matrix. For 2×2: swap diagonal, negate off-diagonal, divide by det. This gives explicit formulas showing how each inverse entry depends on matrix entries.

How is the cross product computed using a determinant?

a × b = det([î, ĵ, k̂; a₁, a₂, a₃; b₁, b₂, b₃]). Expand along row 1: each component is a 2×2 minor. This is a formal determinant (first row has vectors, not numbers) but cofactor expansion works mechanically.

What is the characteristic polynomial?

p(λ) = det(A - λI) is a degree-n polynomial whose roots are the eigenvalues. Key facts: det(A) = product of eigenvalues (set λ=0); trace(A) = sum of eigenvalues (from the λⁿ⁻¹ coefficient).

What is the Wronskian?

The Wronskian W(f₁,...,fₙ) is a determinant with functions in row 1, first derivatives in row 2, etc. If W ≠ 0 at some point, the functions are linearly independent. It's essential for testing solution sets of differential equations.

What is the Vandermonde determinant?

For nodes x₁,...,xₙ, the Vandermonde matrix has entry xᵢʲ⁻¹. Its determinant is ∏(xⱼ - xᵢ) over i < j — nonzero iff all nodes are distinct. This guarantees unique polynomial interpolation through n points.

Determinant Applications: Cramer's Rule, Inverse & More

Cramer's Rule

Given a linear system

Ax = \mathbf{b}

where

A

is

n \times n

with

\det(A) \neq 0

, Cramer's rule expresses each component of the solution directly as a ratio of determinants:

x_i = \frac{\det(A_i)}{\det(A)}

where

A_i

is the matrix formed by replacing column

i

of

A

with the right-hand side vector

\mathbf{b}

. Every other column stays in place.

2×2 Example

For the system

\begin{pmatrix} 3 & 2 \\ 1 & 5 \end{pmatrix} \begin{pmatrix} x_1 \\ x_2 \end{pmatrix} = \begin{pmatrix} 8 \\ 7 \end{pmatrix}

the coefficient determinant is

\det(A) = 3 \cdot 5 - 2 \cdot 1 = 13

. Replacing column

1

with

\mathbf{b}

:

\det(A_1) = \det\begin{pmatrix} 8 & 2 \\ 7 & 5 \end{pmatrix} = 40 - 14 = 26

Replacing column

2

with

\mathbf{b}

:

\det(A_2) = \det\begin{pmatrix} 3 & 8 \\ 1 & 7 \end{pmatrix} = 21 - 8 = 13

So

x_1 = 26/13 = 2

and

x_2 = 13/13 = 1

.

3×3 Example

\begin{pmatrix} 1 & 0 & 2 \\ -1 & 3 & 1 \\ 2 & 1 & 0 \end{pmatrix} \begin{pmatrix} x_1 \\ x_2 \\ x_3 \end{pmatrix} = \begin{pmatrix} 5 \\ 0 \\ 3 \end{pmatrix}

The coefficient determinant is

\det(A) = 1(0 - 1) - 0 + 2(-1 - 6) = -1 - 14 = -15

. The three modified determinants are:

\det(A_1) = \det\begin{pmatrix} 5 & 0 & 2 \\ 0 & 3 & 1 \\ 3 & 1 & 0 \end{pmatrix} = 5(0 - 1) - 0 + 2(0 - 9) = -5 - 18 = -23

\det(A_2) = \det\begin{pmatrix} 1 & 5 & 2 \\ -1 & 0 & 1 \\ 2 & 3 & 0 \end{pmatrix} = 1(0 - 3) - 5(0 - 2) + 2(-3 - 0) = -3 + 10 - 6 = 1

\det(A_3) = \det\begin{pmatrix} 1 & 0 & 5 \\ -1 & 3 & 0 \\ 2 & 1 & 3 \end{pmatrix} = 1(9 - 0) - 0 + 5(-1 - 6) = 9 - 35 = -26

The solution is

x_1 = -23/(-15) = 23/15

,

x_2 = 1/(-15) = -1/15

,

x_3 = -26/(-15) = 26/15

.

Theoretical Significance

Cramer's rule proves that each solution component is a rational function of the matrix entries and the right-hand side entries. This has consequences in pure algebra and in sensitivity analysis, where it shows how solutions respond to perturbations in the data. As a computational method, however, it requires

n + 1

determinant evaluations, making it far more expensive than Gaussian elimination for large systems.

The Inverse via the Adjugate

The adjugate identity

A \cdot \operatorname{adj}(A) = \det(A) \cdot I

immediately gives an explicit formula for the inverse when

\det(A) \neq 0

:

A^{-1} = \frac{1}{\det(A)} \operatorname{adj}(A)

Every entry of

A^{-1}

is expressed as a cofactor of

A

divided by

\det(A)

.

The 2×2 Case

For

A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}

, the cofactor matrix is

\begin{pmatrix} d & -c \\ -b & a \end{pmatrix}

, and transposing gives

\operatorname{adj}(A) = \begin{pmatrix} d & -b \\ -c & a \end{pmatrix}

. The inverse is

A^{-1} = \frac{1}{ad - bc} \begin{pmatrix} d & -b \\ -c & a \end{pmatrix}

This is the familiar swap-the-diagonal, negate-the-off-diagonal formula that appears in every introductory linear algebra course.

3×3 Worked Example

For

A = \begin{pmatrix} 1 & 2 & 0 \\ 0 & 1 & 3 \\ 2 & 0 & 1 \end{pmatrix}

, first compute

\det(A)

by expanding along the first row:

\det(A) = 1(1 - 0) - 2(0 - 6) + 0 = 1 + 12 = 13

The nine cofactors are:

C_{11} = +(1 \cdot 1 - 3 \cdot 0) = 1, \quad C_{12} = -(0 \cdot 1 - 3 \cdot 2) = 6, \quad C_{13} = +(0 \cdot 0 - 1 \cdot 2) = -2

C_{21} = -(2 \cdot 1 - 0 \cdot 0) = -2, \quad C_{22} = +(1 \cdot 1 - 0 \cdot 2) = 1, \quad C_{23} = -(1 \cdot 0 - 2 \cdot 2) = 4

C_{31} = +(2 \cdot 3 - 0 \cdot 1) = 6, \quad C_{32} = -(1 \cdot 3 - 0 \cdot 0) = -3, \quad C_{33} = +(1 \cdot 1 - 2 \cdot 0) = 1

The adjugate is the transpose of the cofactor matrix:

\operatorname{adj}(A) = \begin{pmatrix} 1 & -2 & 6 \\ 6 & 1 & -3 \\ -2 & 4 & 1 \end{pmatrix}

So

A^{-1} = \frac{1}{13} \begin{pmatrix} 1 & -2 & 6 \\ 6 & 1 & -3 \\ -2 & 4 & 1 \end{pmatrix}

.

Verification:

A \cdot A^{-1}

should produce the identity. The

(1,1)

entry is

\frac{1}{13}(1 \cdot 1 + 2 \cdot 6 + 0 \cdot (-2)) = \frac{13}{13} = 1

. The

(1,2)

entry is

\frac{1}{13}(1 \cdot (-2) + 2 \cdot 1 + 0 \cdot 4) = \frac{0}{13} = 0

. The remaining entries check out similarly.

Practical Assessment

The adjugate formula writes every entry of the inverse as an explicit ratio of cofactors and the determinant. This is valuable for symbolic work — it shows exactly how each entry of

A^{-1}

depends on the entries of

A

. For numerical computation on matrices larger than

3 \times 3

, row reduction is vastly more efficient.

The Cross Product as a Determinant

The cross product of two vectors

\mathbf{a} = (a_1, a_2, a_3)

and

\mathbf{b} = (b_1, b_2, b_3)

in

\mathbb{R}^3

can be computed as a symbolic

3 \times 3

determinant:

\mathbf{a} \times \mathbf{b} = \det\begin{pmatrix} \hat{\mathbf{i}} & \hat{\mathbf{j}} & \hat{\mathbf{k}} \\ a_1 & a_2 & a_3 \\ b_1 & b_2 & b_3 \end{pmatrix}

Expanding along the first row using the cofactor formula:

\mathbf{a} \times \mathbf{b} = \hat{\mathbf{i}}(a_2 b_3 - a_3 b_2) - \hat{\mathbf{j}}(a_1 b_3 - a_3 b_1) + \hat{\mathbf{k}}(a_1 b_2 - a_2 b_1)

Each component of the resulting vector is a

2 \times 2

minor — the sub-determinant obtained by deleting the appropriate row and column from the lower two rows.

This is a formal rather than literal use of the determinant. The first row contains basis vectors, not numbers, so the "determinant" is not a scalar but a vector. The cofactor expansion still applies mechanically, and the alternating signs

+, -, +

produce the correct cross product components.

Worked Example

For

\mathbf{a} = (2, -1, 3)

and

\mathbf{b} = (4, 0, -2)

:

\mathbf{a} \times \mathbf{b} = \det\begin{pmatrix} \hat{\mathbf{i}} & \hat{\mathbf{j}} & \hat{\mathbf{k}} \\ 2 & -1 & 3 \\ 4 & 0 & -2 \end{pmatrix}

= \hat{\mathbf{i}}((-1)(-2) - (3)(0)) - \hat{\mathbf{j}}((2)(-2) - (3)(4)) + \hat{\mathbf{k}}((2)(0) - (-1)(4))

= \hat{\mathbf{i}}(2) - \hat{\mathbf{j}}(-16) + \hat{\mathbf{k}}(4) = (2, 16, 4)

The magnitude is

|\mathbf{a} \times \mathbf{b}| = \sqrt{4 + 256 + 16} = \sqrt{276} = 2\sqrt{69}

. This equals the area of the parallelogram spanned by

\mathbf{a}

and

\mathbf{b}

, connecting the cross product back to the geometric interpretation of the determinant as an area measure.

The Characteristic Polynomial

For an

n \times n

matrix

A

, the characteristic polynomial is defined as

p(\lambda) = \det(A - \lambda I)

This is a polynomial of degree

n

in the variable

\lambda

. Its roots are the eigenvalues of

A

— the scalars

\lambda

for which the matrix

A - \lambda I

becomes singular.

2×2 Example

For

A = \begin{pmatrix} 4 & 1 \\ 2 & 3 \end{pmatrix}

:

A - \lambda I = \begin{pmatrix} 4 - \lambda & 1 \\ 2 & 3 - \lambda \end{pmatrix}

p(\lambda) = (4 - \lambda)(3 - \lambda) - 2 = \lambda^2 - 7\lambda + 10 = (\lambda - 2)(\lambda - 5)

The eigenvalues are

\lambda = 2

and

\lambda = 5

.

3×3 Example

For

A = \begin{pmatrix} 2 & 1 & 0 \\ 0 & 3 & 1 \\ 0 & 0 & 1 \end{pmatrix}

:

This is upper triangular, so

A - \lambda I

is also upper triangular with diagonal entries

2 - \lambda

,

3 - \lambda

,

1 - \lambda

. The determinant of a triangular matrix is the product of its diagonal entries:

p(\lambda) = (2 - \lambda)(3 - \lambda)(1 - \lambda)

The eigenvalues are

\lambda = 1, 2, 3

— they sit directly on the diagonal, which is always the case for triangular matrices.

Two Identities

Setting

\lambda = 0

in the characteristic polynomial gives

p(0) = \det(A)

, which means the constant term of the characteristic polynomial is the determinant. Since the roots of

p

are the eigenvalues

\lambda_1, \dots, \lambda_n

, this yields

\det(A) = \lambda_1 \lambda_2 \cdots \lambda_n

The determinant equals the product of all eigenvalues, counted with algebraic multiplicity. A second identity connects the coefficient of

\lambda^{n-1}

to the trace:

\lambda_1 + \lambda_2 + \cdots + \lambda_n = \operatorname{tr}(A)

Together, these two identities link the determinant and trace to the eigenvalue spectrum of the matrix.

The Wronskian

The Wronskian extends the determinant's role as a linear independence test from vectors to functions. Given

n

functions

f_1, f_2, \dots, f_n

, each differentiable at least

n - 1

times, the Wronskian is

W(f_1, \dots, f_n)(x) = \det\begin{pmatrix} f_1(x) & f_2(x) & \cdots & f_n(x) \\ f_1'(x) & f_2'(x) & \cdots & f_n'(x) \\ \vdots & \vdots & \ddots & \vdots \\ f_1^{(n-1)}(x) & f_2^{(n-1)}(x) & \cdots & f_n^{(n-1)}(x) \end{pmatrix}

Each column corresponds to one function, and each row raises the order of differentiation by one. The result is a function of

x

, not a constant.

The Independence Test

If

W(f_1, \dots, f_n)(x_0) \neq 0

at some point

x_0

, then the functions

f_1, \dots, f_n

are linearly independent on any interval containing

x_0

. The logic mirrors the matrix case: a nonzero determinant means the "columns" — here the function-derivative profiles — are not proportional.

The converse requires care. A Wronskian that vanishes everywhere does not automatically imply dependence unless the functions are known to be solutions of a single linear ordinary differential equation. Without that structural assumption, counterexamples exist.

Worked Example

Take

f_1 = e^x

,

f_2 = e^{2x}

,

f_3 = e^{3x}

. The Wronskian matrix is

\begin{pmatrix} e^x & e^{2x} & e^{3x} \\ e^x & 2e^{2x} & 3e^{3x} \\ e^x & 4e^{2x} & 9e^{3x} \end{pmatrix}

Factoring

e^x

from column

1

,

e^{2x}

from column

2

, and

e^{3x}

from column

3

:

W = e^{6x} \det\begin{pmatrix} 1 & 1 & 1 \\ 1 & 2 & 3 \\ 1 & 4 & 9 \end{pmatrix}

The remaining matrix is a Vandermonde matrix with nodes

1, 2, 3

. Its determinant is

(2 - 1)(3 - 1)(3 - 2) = 1 \cdot 2 \cdot 1 = 2

. So

W = 2e^{6x}

, which is nonzero for all

x

, confirming that

e^x, e^{2x}, e^{3x}

are linearly independent.

Context

The Wronskian arises most naturally in the theory of linear ordinary differential equations, where it determines whether a proposed set of solutions forms a fundamental system. Abel's identity gives a differential equation for the Wronskian itself, relating its evolution to the coefficient in the ODE. These developments belong to differential equations rather than linear algebra, but the underlying mechanism — testing independence via a determinant — is purely algebraic.

Vandermonde and Structured Determinants

Certain matrices with patterned entries have determinants that admit elegant closed-form expressions. The most important of these is the Vandermonde matrix.

The Vandermonde Determinant

An

n \times n

Vandermonde matrix is built from

n

distinct nodes

x_1, x_2, \dots, x_n

:

V = \begin{pmatrix} 1 & x_1 & x_1^2 & \cdots & x_1^{n-1} \\ 1 & x_2 & x_2^2 & \cdots & x_2^{n-1} \\ \vdots & \vdots & \vdots & \ddots & \vdots \\ 1 & x_n & x_n^2 & \cdots & x_n^{n-1} \end{pmatrix}

Its determinant has the closed form

\det(V) = \prod_{1 \leq i < j \leq n} (x_j - x_i)

The product runs over all pairs with

j > i

, so it contains

\binom{n}{2}

factors. Each factor is a difference between two nodes.

3×3 Verification

For nodes

x_1 = 1

,

x_2 = 2

,

x_3 = 4

:

V = \begin{pmatrix} 1 & 1 & 1 \\ 1 & 2 & 4 \\ 1 & 4 & 16 \end{pmatrix}

Direct expansion:

\det(V) = 1(32 - 16) - 1(16 - 4) + 1(4 - 2) = 16 - 12 + 2 = 6

.

The product formula:

(x_2 - x_1)(x_3 - x_1)(x_3 - x_2) = (2 - 1)(4 - 1)(4 - 2) = 1 \cdot 3 \cdot 2 = 6

.

Why It Matters

The Vandermonde determinant is nonzero precisely when all nodes are distinct. This guarantees that a polynomial of degree at most

n - 1

is uniquely determined by its values at

n

distinct points — the theoretical foundation of polynomial interpolation. It also appears in the theory of symmetric polynomials and in the derivation of various discrete orthogonality relations.

Other Structured Determinants

Several other matrix families have known determinant formulas. Circulant matrices, built from cyclic shifts of a single row, have determinants expressible through the discrete Fourier transform: if the first row is

(c_0, c_1, \dots, c_{n-1})

, then

\det(C) = \prod_{k=0}^{n-1} p(\omega^k)

where

p(x) = c_0 + c_1 x + \cdots + c_{n-1} x^{n-1}

and

\omega = e^{2\pi i/n}

is a primitive

n

-th root of unity.

Hilbert matrices, with entries

H_{ij} = \frac{1}{i + j - 1}

, have a closed-form determinant involving products of factorials. These matrices are notoriously ill-conditioned — their determinants shrink rapidly as

n

grows, reflecting extreme sensitivity to perturbation.

Tridiagonal matrices, with nonzero entries only on the main diagonal and the two adjacent diagonals, have determinants satisfying a three-term recurrence: if

D_n

denotes the determinant of the

n \times n

tridiagonal matrix, then

D_n = a_n D_{n-1} - b_n c_n D_{n-2}

, where

a_n

is the

n

-th diagonal entry and

b_n, c_n

are the adjacent off-diagonal entries. This recurrence allows

O(n)

computation, much faster than general methods.

Each of these families illustrates the same principle: when a matrix has special structure, its determinant often has a formula that exploits that structure directly, bypassing both cofactor expansion and row reduction.

Applications of Determinants

Solving Problems with Determinants

Cramer's Rule

2×2 Example

3×3 Example

Theoretical Significance

The Inverse via the Adjugate

The 2×2 Case

3×3 Worked Example

Practical Assessment

The Cross Product as a Determinant

Worked Example

The Characteristic Polynomial

2×2 Example

3×3 Example

Two Identities

The Wronskian

The Independence Test

Worked Example

Context

Vandermonde and Structured Determinants

The Vandermonde Determinant

3×3 Verification

Why It Matters

Other Structured Determinants