Matrices — Operations and Determinants

⚡ Key Takeaways

A matrix is a rectangular array of numbers; matrix operations (addition, scalar multiplication, multiplication, transposition, inversion) combine matrices, and the determinant is a single scalar computed from a square matrix that measures invertibility and signed volume scaling. Key result: a square matrix is invertible if and only if its determinant is nonzero. Applies to: solving linear systems, transformations of the plane and space, computer graphics, statistics. Who/when: developed by Cayley, Sylvester, Cramer, Leibniz, and Laplace from the 17th to 19th centuries.

💡 The Core Insight

A matrix is a linear transformation in disguise. Every operation on matrices corresponds to an operation on transformations: addition combines transformations, multiplication composes them (do one, then the other), and the determinant measures how much the transformation stretches or flips space. When the determinant is zero, the transformation crushes space into a lower dimension and cannot be reversed — that single number captures the entire question of solvability.

DefinitionMatrix and Basic Notation

An $m \times n$ matrix $A$ over the real numbers is an array of $mn$ entries arranged in $m$ rows and $n$ columns:

1

\quad A = \begin{bmatrix} a_{11} & a_{12} & \cdots & a_{1n} \\ a_{21} & a_{22} & \cdots & a_{2n} \\ \vdots & \vdots & \ddots & \vdots \\ a_{m1} & a_{m2} & \cdots & a_{mn} \end{bmatrix}

Derivation hint: this is a definition; $a_{ij}$ denotes the entry in row $i$ ( $1 \le i \le m$ ) and column $j$ ( $1 \le j \le n$ ). A matrix with $m = n$ is square. The main diagonal consists of entries $a_{ii}$ . The identity matrix $I_n$ has $a_{ii}=1$ and $a_{ij}=0$ for $i \ne j$ .

DefinitionAddition and Scalar Multiplication

For two matrices $A=[a_{ij}]$ and $B=[b_{ij}]$ of the same size $m \times n$ , and a scalar $c \in \mathbb{R}$ :

2

\quad (A+B)_{ij} = a_{ij} + b_{ij}, \qquad (cA)_{ij} = c\,a_{ij}

Derivation hint: addition and scaling act entrywise; the sizes must match for addition.

DefinitionMatrix Multiplication

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

3

\quad (AB)_{ij} = \sum_{k=1}^{n} a_{ik}\, b_{kj}

Derivation hint: entry $(i,j)$ is the dot product of row $i$ of $A$ with column $j$ of $B$ .

DefinitionTranspose

The transpose $A^{T}$ of an $m \times n$ matrix $A$ is the $n \times m$ matrix obtained by swapping rows and columns:

4

\quad (A^{T})_{ij} = a_{ji}

Derivation hint: reflect entries across the main diagonal; row $i$ becomes column $i$ .

DefinitionDeterminant

The determinant $\det(A)$ of a square $n\times n$ matrix is a scalar. For $2\times2$ and $3\times3$ matrices:

5

\quad \det\begin{bmatrix} a & b \\ c & d \end{bmatrix} = ad - bc

Derivation hint: signed area of the parallelogram spanned by the rows $(a,b)$ and $(c,d)$ .

6

\quad \det\begin{bmatrix} a & b & c \\ d & e & f \\ g & h & i \end{bmatrix} = a(ei-fh) - b(di-fg) + c(dh-eg)

Derivation hint: cofactor (Laplace) expansion along the first row; alternating signs $+,-,+$ . In general, the cofactor expansion along row $i$ is:

7

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

Derivation hint: $M_{ij}$ is the <strong>minor</strong> — the determinant of the $(n-1)\times(n-1)$ matrix formed by deleting row $i$ and column $j$ .

TheoremInvertibility Criterion

A square matrix $A$ has an inverse $A^{-1}$ satisfying $AA^{-1} = A^{-1}A = I$ if and only if $\det(A) \ne 0$ . When it exists, for the $2\times2$ case:

8

\quad A^{-1} = \frac{1}{\det(A)}\begin{bmatrix} d & -b \\ -c & a \end{bmatrix}, \quad A=\begin{bmatrix} a & b \\ c & d\end{bmatrix}

Derivation hint: swap diagonal entries, negate off-diagonal entries, divide by the determinant; verify by direct multiplication that $AA^{-1}=I$ .

TheoremKey Determinant Properties

For square $n\times n$ matrices $A$ , $B$ :

9

\quad \det(AB) = \det(A)\det(B), \qquad \det(A^{T}) = \det(A), \qquad \det(I_n)=1

Derivation hint: multiplicativity follows from the volume-scaling interpretation — composing transformations multiplies their scaling factors.

ProofThe 2×2 Inverse Formula

We verify equation [8] directly by multiplication. 1. Let $A = \begin{bmatrix} a & b \\ c & d\end{bmatrix}$ with $\det(A) = ad - bc \ne 0$ . 2. Define the candidate $B = \dfrac{1}{ad-bc}\begin{bmatrix} d & -b \\ -c & a\end{bmatrix}$ . 3. Compute $AB$ using equation [3]:

AB = \frac{1}{ad-bc}\begin{bmatrix} a & b \\ c & d\end{bmatrix}\begin{bmatrix} d & -b \\ -c & a\end{bmatrix} = \frac{1}{ad-bc}\begin{bmatrix} ad-bc & -ab+ab \\ cd-cd & -cb+ad \end{bmatrix}.

4. Simplify: the off-diagonal terms vanish, the diagonal terms equal $ad-bc$ :

AB = \frac{1}{ad-bc}\begin{bmatrix} ad-bc & 0 \\ 0 & ad-bc \end{bmatrix} = \begin{bmatrix} 1 & 0 \\ 0 & 1\end{bmatrix} = I.

5. A symmetric computation gives $BA = I$ . Hence $B = A^{-1}$ . 6. If $\det(A)=0$ , division is impossible and no inverse exists — confirming the criterion. $\blacksquare$

∎

ProofMultiplicativity Sketch: det(AB) = det(A)det(B)

1. Interpret a matrix as a linear transformation of $\mathbb{R}^n$ ; $\det$ is the signed factor by which it scales $n$ -dimensional volume. 2. The product $AB$ represents applying $B$ first, then $A$ (composition). 3. Volume scaling under composition multiplies: scale by $\det(B)$ , then by $\det(A)$ . 4. Therefore the total scaling is $\det(A)\det(B)$ , giving $\det(AB)=\det(A)\det(B)$ . $\blacksquare$

∎

mathematician · Arthur Cayley | 1821–1895

The English mathematician Arthur Cayley is credited with founding the systematic theory of matrices in his 1858 Memoir on the Theory of Matrices. He introduced matrix multiplication as we use it today, defined the identity and inverse, and recognized matrices as objects with their own algebra — not merely shorthand for systems of equations. He stated (and verified in low dimensions) what is now the Cayley–Hamilton theorem: every square matrix satisfies its own characteristic equation. Working closely with his friend James Joseph Sylvester (who coined the word "matrix" in 1850), Cayley transformed scattered computational tricks into a unified abstract subject.

🧠 Intuition

Think of a $2\times2$ matrix as a machine that bends the plane. The columns of $A$ tell you where the basis arrows $(1,0)$ and $(0,1)$ land after the transformation. The determinant is the area of the parallelogram those two landing-arrows span — and its sign tells you whether orientation was preserved (positive) or flipped like a mirror (negative). If the two columns point along the same line, the parallelogram collapses to a sliver of zero area: the determinant is $0$ , the plane is squashed onto a line, and information is lost forever — that is exactly why you cannot invert. Multiplication composes machines: $AB$ means "first run $B$ , then run $A$ ," which is why order matters and why scaling factors multiply.

Figure

Geometric meaning of the 2×2 determinant: the area of the parallelogram spanned by the matrix's columns

ExampleMultiplying Two Matrices

Given: $A = \begin{bmatrix} 1 & 2 \\ 3 & 4 \end{bmatrix}$ ( $2\times2$ ), $B = \begin{bmatrix} 5 & 6 \\ 7 & 8 \end{bmatrix}$ ( $2\times2$ ). Find $AB$ .

Steps (apply equation [3], dot rows of $A$ with columns of $B$ ):

$(AB)_{11} = 1\cdot5 + 2\cdot7 = 5 + 14 = 19$
$(AB)_{12} = 1\cdot6 + 2\cdot8 = 6 + 16 = 22$
$(AB)_{21} = 3\cdot5 + 4\cdot7 = 15 + 28 = 43$
$(AB)_{22} = 3\cdot6 + 4\cdot8 = 18 + 32 = 50$

Check non-commutativity: $BA = \begin{bmatrix} 23 & 34 \\ 31 & 46\end{bmatrix} \ne AB$ .

\boxed{AB = \begin{bmatrix} 19 & 22 \\ 43 & 50 \end{bmatrix}}

ExampleDeterminant and Inverse, with a System Solved by Cramer's Rule

Given: Solve $\begin{cases} 2x + 3y = 8 \\ 1x + 4y = 9 \end{cases}$ , i.e. $A\mathbf{x}=\mathbf{b}$ with $A=\begin{bmatrix} 2 & 3 \\ 1 & 4\end{bmatrix}$ , $\mathbf{b}=\begin{bmatrix} 8 \\ 9\end{bmatrix}$ .

Step 1 — Determinant (equation [5]): $\det(A) = 2\cdot4 - 3\cdot1 = 8 - 3 = 5 \ne 0$ , so a unique solution exists.

Step 2 — Cramer's Rule. Replace column 1 with $\mathbf{b}$ for $x$ , column 2 for $y$ :

x = \frac{\det\begin{bmatrix} 8 & 3 \\ 9 & 4\end{bmatrix}}{5} = \frac{8\cdot4 - 3\cdot9}{5} = \frac{32-27}{5} = \frac{5}{5} = 1.

y = \frac{\det\begin{bmatrix} 2 & 8 \\ 1 & 9\end{bmatrix}}{5} = \frac{2\cdot9 - 8\cdot1}{5} = \frac{18-8}{5} = \frac{10}{5} = 2.

Step 3 — Verify with the original equations: $2(1)+3(2)=8$ ✓, $1(1)+4(2)=9$ ✓.

\boxed{x = 1,\quad y = 2}

⚠️ Common mistake

Matrix multiplication is not commutative. In general $AB \ne BA$ , and one product may be defined while the other is not (dimensions). Also: $AB = 0$ does not imply $A=0$ or $B=0$ — there are nonzero "zero divisors." And there is no "matrix division"; you multiply by the inverse $A^{-1}$ (when it exists), being careful about whether you multiply on the left ( $A^{-1}B$ ) or the right ( $BA^{-1}$ ), since these differ.

Counterexample

Let $A = \begin{bmatrix} 0 & 1 \\ 0 & 0\end{bmatrix}$ . Then $A \ne 0$ , yet $A^2 = \begin{bmatrix} 0 & 1 \\ 0 & 0\end{bmatrix}\begin{bmatrix} 0 & 1 \\ 0 & 0\end{bmatrix} = \begin{bmatrix} 0 & 0 \\ 0 & 0\end{bmatrix} = 0$ . A nonzero matrix can square to the zero matrix — impossible for ordinary nonzero numbers. Note $\det(A)=0\cdot0-1\cdot0=0$ , so $A$ is non-invertible, consistent with Theorem (Invertibility Criterion).

connection

Matrices are the computational engine behind several neighbouring topics. Solving $A\mathbf{x}=\mathbf{b}$ links directly to Systems of Linear Equations and Gaussian Elimination. The determinant's role as a signed area/volume connects to Vectors and the Cross Product (the $3\times3$ determinant formula for the cross product) and to Linear Transformations. Eigenvalues, found by solving $\det(A-\lambda I)=0$ , lead into Eigenvalues and Eigenvectors. The transpose and dot product underpin Inner Product Spaces, and matrix powers connect to Markov Chains.

🔬 Where this appears

⏳ History

- c. 200 BCE — The Chinese text Nine Chapters on the Mathematical Art solves systems of linear equations using array (matrix-like) elimination, predating Western methods by centuries.

- 1683 — Seki Takakazu in Japan and 1693 — Gottfried Leibniz in Europe independently introduce determinants while studying systems of equations.

- 1750 — Gabriel Cramer publishes Cramer's Rule, expressing solutions of linear systems as ratios of determinants.

- 1812 — Augustin-Louis Cauchy proves the multiplication theorem

\det(AB)=\det(A)\det(B)

and gives the modern definition of determinant.

- 1850 — James Joseph Sylvester coins the word "matrix" (Latin for "womb"), viewing it as a source from which determinants are born.

- 1858 — Arthur Cayley's Memoir on the Theory of Matrices establishes matrix algebra, multiplication, inverses, and the identity matrix.

- 1878 — Ferdinand Georg Frobenius unifies and extends the theory, developing rank, eigenvalues, and the characteristic polynomial.

- 20th century — Matrices become foundational to quantum mechanics (Heisenberg's 1925 matrix mechanics), numerical computing, and data science.

open-problem

While matrix theory is fully understood mathematically, efficient computation remains an active research frontier. Naïve matrix multiplication of two $n\times n$ matrices costs $O(n^3)$ operations; Strassen (1969) achieved $O(n^{2.807})$ , and the current record is roughly $O(n^{2.371})$ — but whether multiplication can be done in essentially $O(n^2)$ time (the matrix multiplication exponent $\omega = 2$ question) is unknown. Likewise, computing large determinants is numerically delicate: cofactor expansion is hopelessly slow ( $n!$ terms), and floating-point cancellation can make $\det$ near zero indistinguishable from zero, so in practice the determinant is rarely used directly to test invertibility — conditioning and LU decomposition are used instead.

Matrices — Operations and Determinants

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also