Matrices (Precalculus)

A matrix is a rectangular array of numbers arranged in rows and columns. Matrices encode linear systems, transformations, and data tables; they can be added, scaled, multiplied, inverted, and reduced. Used everywhere from solving systems of equations to computer graphics, economics, and statistics. Formalized in the 19th century by Cayley, Sylvester, and others.

A matrix is a rectangular array of entries (usually real numbers) organized into $m$ rows and $n$ columns, written as an $m \times n$ object. Matrices matter because they package an entire system of linear equations into a single algebraic object, allowing systematic solution by row operations and giving a compact language for linear transformations. They support a rich arithmetic — addition, scalar multiplication, and a non-commutative multiplication — that obeys most familiar algebra laws but breaks a few crucial ones. The surprising fact: matrix multiplication is not commutative ( $A B \neq = B A$ in general), and some nonzero matrices have no inverse — phenomena with no analogue in ordinary number arithmetic. This single tool unifies geometry, algebra, and data.

The Core Insight

A matrix is simultaneously three things: (1) a bookkeeping table of numbers, (2) a compact encoding of a linear system $A x = b$ , and (3) a machine that transforms vectors by multiplication. Mastering matrices means fluidly switching among these three viewpoints. The operations of row reduction solve systems; the operation of multiplication composes transformations; the inverse undoes them.

Formal Statement

Matrix and Its Dimensions

An $m \times n$ matrix $A$ is a rectangular array of numbers with $m$ rows and $n$ columns:

A = a_{11} a_{21} ⋮ a_{m 1} a_{12} a_{22} ⋮ a_{m 2} \dots \dots ⋱ \dots a_{1 n} a_{2 n} ⋮ a_{mn}

⟳The entry

a_{ij}

sits in row

i

(counted top to bottom) and column

j

(counted left to right); the pair

(m, n)

is the matrix's

order or dimension.

Here $a_{ij}$ denotes the entry in row $i$ , column $j$ . If $m = n$ the matrix is square of order $n$ . A $1 \times n$ matrix is a row vector; an $m \times 1$ matrix is a column vector.

Matrix Addition and Scalar Multiplication

For two matrices $A$ and $B$ of the same order $m \times n$ , and a scalar $c$ :

(A + B)_{ij} = a_{ij} + b_{ij}, (c A)_{ij} = c a_{ij}

⟳Both operations act

entry by entry; addition requires matching dimensions, scalar multiplication does not.

Matrix Multiplication

If $A$ is $m \times n$ and $B$ is $n \times p$ (the inner dimensions must match), the product $A B$ is $m \times p$ with entries:

(A B)_{ik} = j = 1 \sum n a_{ij} b_{j k}

⟳Entry

(i, k)

is the

dot product of row $i$ of $A$ with column $k$ of $B$ ; this is the rule that makes matrix multiplication encode composition of linear transformations.

Identity and Inverse

The $n \times n$ identity matrix $I_{n}$ has $1$ s on the main diagonal and $0$ s elsewhere. A square matrix $A$ is invertible if there exists $A^{- 1}$ with:

A A^{- 1} = A^{- 1} A = I_{n}

⟳

I_{n}

acts as the multiplicative "

1

"; the inverse exists exactly when

det A \neq = 0

Determinant of a 2×2 Matrix and Its Inverse

For $A = [a c b d]$ , the determinant is $det A = a d - b c$ , and when $det A \neq = 0$ :

A^{- 1} = \frac{1}{a d - b c} [d - c - b a]

⟳Swap the diagonal entries, negate the off-diagonal entries, and divide by the determinant; verify by computing

A A^{- 1} = I

Why It Holds / Derivation

The 2×2 Inverse Formula

We seek $A^{- 1} = [x z y w]$ such that $A A^{- 1} = I$ .

Step 1. Write the matrix equation:

[a c b d] [x z y w] = [1001]

⟳Multiply using rule [3] to get four scalar equations.

Step 2. The four resulting equations are

a x + b z = 1

a y + b w = 0

c x + d z = 0

cy + d w = 1

. Solving the first column system

{a x + b z = 1, c x + d z = 0}

x = \frac{d}{a d - b c}, z = \frac{- c}{a d - b c}

⟳Eliminate by cross-multiplication; the common denominator

a d - b c

emerges naturally.

Step 3. Solving the second column system

{a y + b w = 0, cy + d w = 1}

y = \frac{- b}{a d - b c}, w = \frac{a}{a d - b c}

⟳Same elimination procedure applied to the right-hand column of

I

Step 4. Assembling

x, y, z, w

reproduces formula [5]. The procedure divides by

a d - b c

, which fails exactly when

det A = 0

— proving invertibility requires a nonzero determinant.

■

Arthur Cayley | 1821–1895

The English mathematician Arthur Cayley is regarded as the founder of matrix theory as an abstract subject. In his 1858 Memoir on the Theory of Matrices he treated matrices as objects in their own right, defining their addition, multiplication, and inverses, and stating (for $2 \times 2$ and $3 \times 3$ cases) what became the Cayley–Hamilton theorem: every square matrix satisfies its own characteristic equation. A barrister by profession for fourteen years, Cayley published nearly a thousand papers. His collaborator James Joseph Sylvester coined the word "matrix" in 1850, picturing it as the "womb" (Latin matrix) from which determinants are born.

Building Intuition

Think of a matrix three ways at once. As a table: a spreadsheet of numbers — sales by month, pixel brightnesses, coefficients of equations. As a system solver: the array $[21 3 - 1]$ paired with right-hand side $[81]$ is the system $2 x + 3 y = 8, x - y = 1$ . As a transformer: multiplying a vector by a matrix stretches, rotates, shears, or projects it. The reason multiplication isn't commutative becomes obvious here: rotating then reflecting is generally not the same as reflecting then rotating. The determinant measures how much the transformation scales areas; when it's zero, the transformation collapses space onto a line (or point), destroying information, which is exactly why you can't invert it.

A 2×2 matrix transforming the unit square into a parallelogram; the determinant equals the new area.

Worked Examples

Solving a Linear System with the Inverse

Given: Solve

{2 x + 3 y = 8 x - y = 1

using a matrix inverse. Step 1. Write as

A x = b

with

A = [21 3 - 1]

b = [81]

. Step 2. Compute

det A = (2) (- 1) - (3) (1) = - 2 - 3 = - 5 \neq = 0

, so

A^{- 1}

exists. Step 3. By formula [5]:

A^{- 1} = \frac{1}{- 5} [- 1 - 1 - 3 2] = [1/5 1/5 3/5 - 2/5]

Step 4. Multiply:

x = A^{- 1} b = [1/5 1/5 3/5 - 2/5] [81] = [8/5 + 3/5 8/5 - 2/5] = [11/5 6/5]

x = \frac{11}{5} = 2.2, y = \frac{6}{5} = 1.2

Check:

2 (2.2) + 3 (1.2) = 4.4 + 3.6 = 8

✓ and

2.2 - 1.2 = 1

✓.

Matrix Multiplication and Non-Commutativity

Given:

A = [1021]

B = [1301]

. Compute

A B

and

B A

. Step 1. Compute

A B

using rule [3], row-by-column:

A B = [(1) (1) + (2) (3) (0) (1) + (1) (3) (1) (0) + (2) (1) (0) (0) + (1) (1)] = [7321]

Step 2. Compute

B A

B A = [(1) (1) + (0) (0) (3) (1) + (1) (0) (1) (2) + (0) (1) (3) (2) + (1) (1)] = [1327]

Step 3. Compare:

A B = [7321] \neq = [1327] = B A

This concretely demonstrates that matrix multiplication is not commutative.

Common Misconceptions

Multiplication Pitfalls

Dimensions must align. You can only form $A B$ when the number of columns of $A$ equals the number of rows of $B$ . A $2 \times 3$ times a $2 \times 3$ is undefined.
Order matters. $A B \neq = B A$ in general; never silently swap factors.
No automatic cancellation. $A B = A C$ does not imply $B = C$ unless $A$ is invertible.
There is no matrix division. "Dividing by $A$ " means multiplying by $A^{- 1}$ — and you must choose a side: $A^{- 1} B \neq = B A^{- 1}$ in general.
Determinant zero blocks inversion. If $det A = 0$ , the matrix is singular and has no inverse.

Nonzero Matrices with Zero Product

With ordinary numbers, $x y = 0$ forces $x = 0$ or $y = 0$ . Matrices break this. Let

A = [1000], B = [0001] .

Then

A B = [0000]

, the zero matrix, even though neither

A

nor

B

is zero. Such matrices are zero divisors; both are singular (

det = 0

), consistent with the warning above.

Connections

Matrices sit at the crossroads of several precalculus topics. They are the natural tool for Systems of Linear Equations, converting elimination into mechanical Row Reduction (Gaussian Elimination). The Determinant generalizes the $2 \times 2$ formula and governs invertibility and area/volume scaling. Multiplication of matrices is exactly composition of Linear Transformations, linking matrices to Vectors and geometry. The numerical recipe Cramer's Rule expresses solutions via determinants. At a deeper level, matrices launch the entire subject of Linear Algebra, including eigenvalues and vector spaces.

Real-World Applications

Computer graphics & game engines. Rotation, scaling, translation, and perspective of 3D scenes are all matrix multiplications applied to vertex coordinates (the graphics pipeline uses $4 \times 4$ matrices).
Economics — Leontief input–output models. A matrix encodes how much of each industry's output every other industry consumes; solving $(I - A) x = d$ predicts production needed to meet demand.
Google's PageRank. Web pages and links form a giant matrix; the dominant eigenvector ranks pages by importance.
Cryptography — Hill cipher. Blocks of plaintext are multiplied by an invertible matrix mod 26; decryption uses the inverse matrix.
Markov chains & weather/finance models. A transition matrix of probabilities, multiplied repeatedly, predicts long-run state distributions.
Computer vision & machine learning. Images are matrices of pixel values; neural networks are stacked matrix multiplications with nonlinear activations.
Engineering — structural analysis. Stiffness matrices relate forces and displacements in trusses and circuits via $K u = f$ .

Historical Development

c. 200 BCE — China. The Nine Chapters on the Mathematical Art solves systems by arranging coefficients in a grid and performing column operations — essentially Gaussian elimination two millennia early.
1683 — Japan & Europe. Seki Takakazu in Japan and (1693) Leibniz in Europe independently develop determinants while studying systems of equations.
1750 — Cramer. Gabriel Cramer publishes the determinant-based rule now bearing his name for solving linear systems.
1810s — Gauss. Carl Friedrich Gauss systematizes elimination while computing orbits and least-squares fits.
1850 — Sylvester. James Joseph Sylvester coins the term "matrix."
1858 — Cayley. Arthur Cayley's Memoir on the Theory of Matrices establishes matrix algebra as an independent subject and states the Cayley–Hamilton theorem.
1888 — Peano. Giuseppe Peano gives the axiomatic definition of a vector space, providing the abstract home for matrices.
1920s — Quantum mechanics. Heisenberg, Born, and Jordan formulate matrix mechanics, revealing that non-commutativity ( $P Q - QP \neq = 0$ ) is physically fundamental.
1947 onward — Computing era. Von Neumann, Turing, and others develop numerical matrix algorithms; LU decomposition, eigenvalue methods, and the digital computer make large-scale matrix computation routine.

Subtleties & Limits

At the precalculus level, three subtleties point toward deeper waters. (1) Numerical stability: computing inverses by the cofactor formula is hopeless for large matrices; real software uses factorizations (LU, QR), and near-singular matrices amplify rounding error — quantified by the condition number. (2) Beyond $2 \times 2$ and $3 \times 3$ : the determinant for an $n \times n$ matrix has $n!$ terms, so naive evaluation is computationally explosive; efficient algorithms are a serious topic. (3) The structure of singular systems: when $det A = 0$ , a system may have no solution or infinitely many — distinguishing these requires the concept of rank, the gateway to vector spaces and dimension. Open research areas include fast matrix multiplication (the true exponent below $n^{2.37}$ is unknown) and the numerical analysis of huge sparse matrices arising in scientific computing.