Matrix Representation — Matrix of a Linear Transformation |...

SECTION 1

Historical Context & Motivation

Imagine you have a machine that takes in a point on a grid and spits out a new point somewhere else. That machine might rotate every point, stretch them, flip them, or squish them together. Mathematicians call such a machine a linear transformation. For centuries, people described these transformations with words and pictures. The breakthrough came when mathematicians realized they could encode every linear transformation as a neat grid of numbers — a matrix. This idea turned geometry into arithmetic and opened the door to modern computing, graphics, and data science.

1693

Leibniz's Determinant Idea

Gottfried Wilhelm Leibniz wrote about systems of equations and hinted at organizing coefficients into rectangular arrays, planting the earliest seeds of matrix thinking.

1826

Cauchy Formalizes Determinants

Augustin-Louis Cauchy developed a systematic theory of determinants, bringing rigor to the algebra of rectangular arrays of numbers.

1858

Cayley Defines the Matrix

Arthur Cayley published the first formal definition of a matrix and rules for matrix multiplication, establishing the language we still use today.

1888

Peano Defines Vector Spaces

Giuseppe Peano gave the first abstract definition of a vector space, creating the setting in which linear transformations and their matrices live.

1950s+

Computers & Matrices Everywhere

Digital computers made matrix calculations fast and practical. Today matrices power search engines, video games, machine learning, and much more.

The central question this lesson answers is: How do we translate a geometric action — like rotating or stretching — into a compact grid of numbers that a computer (or you!) can work with?

SECTION 2

Core Principles & Definitions

Before we build the matrix, let's nail down the key ideas. A linear transformation is a special kind of function that maps vectors to vectors while obeying two rules: it preserves vector addition and scalar multiplication. In everyday terms, straight lines stay straight and the origin stays fixed. The magic is that we can describe any such transformation completely by knowing what it does to just the basis vectors — the simplest building blocks of our coordinate system.

Linear Transformation

A function T that takes a vector and outputs a new vector, while satisfying T(u + v) = T(u) + T(v) and T(cv) = cT(v).

Standard Basis Vectors

In 2D, the standard basis vectors are e₁ = (1, 0) pointing right and e₂ = (0, 1) pointing up. Every other vector is a combination of these two.

Matrix of a Transformation

Apply the transformation T to each basis vector. Write each result as a column. The resulting grid of numbers is the matrix that represents T.

Matrix–Vector Multiplication

Multiplying the matrix by any input vector gives you exactly the same output as applying the transformation T directly. The matrix is the transformation.

✦ KEY TAKEAWAY

Think of a matrix like a recipe card for a machine. If you know what the machine does to each basic ingredient (each basis vector), you can predict what it does to any mixture of those ingredients. The matrix is just a compact way to write that recipe down.

SECTION 3

Visual Explanation

The diagram below shows how a linear transformation acts on the two standard basis vectors in 2D. On the left you see the original grid with e₁ (the red arrow) and e₂ (the blue arrow). On the right you see where the transformation sends them. The columns of the matrix are exactly the landing spots of e₁ and e₂.

The red arrow e₁ lands at (2, 1) and the blue arrow e₂ lands at (−1, 2). These landing spots become the columns of the matrix A.

Notice the pattern: the first column of the matrix is the output of T applied to e₁, and the second column is the output of T applied to e₂. This works in any dimension — in 3D you would have three basis vectors and three columns. Once you have the matrix, you can compute T of any vector just by multiplying.

SECTION 4

Mathematical Framework

Let's put the recipe into precise math. Suppose T is a linear transformation from ℝⁿ to ℝ^m. The standard matrix A of T is built by placing the images of the standard basis vectors as columns.

STANDARD MATRIX CONSTRUCTION

A = [ T(e₁) T(e₂) ··· T(eₙ) ]

Each T(eⱼ) is an m-dimensional column vector. A is an m × n matrix.

TRANSFORMATION AS MULTIPLICATION

T(x) = A · x

For any vector x in ℝⁿ, the transformation output equals the matrix A multiplied by x. This turns a geometric action into simple arithmetic.

Let's see why this works. Any vector x in ℝⁿ can be written as x = x₁e₁ + x₂e₂ + ··· + xₙeₙ. Because T is linear, we can distribute T across each piece:

LINEARITY IN ACTION

T(x) = x₁ · T(e₁) + x₂ · T(e₂) + ··· + xₙ · T(eₙ)

This sum of scaled columns is exactly what matrix–vector multiplication computes. That's the whole proof!

💡 Quick Check

In 2D, the standard matrix A is 2 × 2. In 3D, if T maps ℝ³ to ℝ³, then A is 3 × 3. If T maps ℝ² to ℝ³ (turning 2D vectors into 3D ones), then A is 3 × 2. The number of rows equals the output dimension and the number of columns equals the input dimension.

SECTION 5

Common Linear Transformations & Their Matrices

Different geometric actions produce different matrices. The diagram below shows four of the most common 2D linear transformations — rotation, reflection, scaling, and shear — along with their standard matrices. Study each one by looking at where e₁ and e₂ land.

Four common 2D transformations. Solid arrows show the original basis vectors; dashed arrows show where they land. Each matrix is built by reading off the new positions of e₁ and e₂ as columns.

Standard matrices for common 2D transformations

Transformation	What It Does	Matrix (2D)
90° Rotation	Spins every vector 90° counterclockwise around the origin	`[0 −1; 1 0]`
Reflection (x-axis)	Flips every vector over the horizontal axis	`[1 0; 0 −1]`
Scaling	Stretches or compresses along each axis independently	`[sₓ 0; 0 sᵧ]`
Shear	Tilts one axis while keeping the other fixed, like leaning a stack of books	`[1 k; 0 1]`

SECTION 6

Worked Example

Let's walk through a complete example. Suppose a linear transformation T maps 2D vectors to 2D vectors, and we know that T(e₁) = (3, 1) and T(e₂) = (−2, 4). We want to find the matrix A and then use it to compute T(5, 2).

Finding and Using the Matrix of T

Step 1 — Identify the Images of the Basis VectorsWe are given that T(e₁) = (3, 1) and T(e₂) = (−2, 4). These two outputs are all we need to build the matrix.

Step 2 — Build the Matrix APlace T(e₁) as the first column and T(e₂) as the second column:

A = [ 3 −2 ] / [ 1 4 ] (reading row by row: first row is 3, −2; second row is 1, 4)

Step 3 — Set Up the MultiplicationTo find T(5, 2), we compute A · x where x = (5, 2). Matrix–vector multiplication gives us: first component = (row 1) · x = 3 × 5 + (−2) × 2, and second component = (row 2) · x = 1 × 5 + 4 × 2.

Step 4 — Compute Each ComponentFirst component: 3 × 5 + (−2) × 2 = 15 − 4 = 11. Second component: 1 × 5 + 4 × 2 = 5 + 8 = 13.

Step 5 — State the AnswerTherefore:

T(5, 2) = (11, 13)

🔍 Why It Works

Think of (5, 2) as "5 copies of e₁ plus 2 copies of e₂." Linearity tells us T(5, 2) = 5 · T(e₁) + 2 · T(e₂) = 5 · (3, 1) + 2 · (−2, 4) = (15, 5) + (−4, 8) = (11, 13). That's the same answer!

SECTION 7

Strengths & Limitations

Matrix representation is one of the most powerful tools in linear algebra, but like any tool, it has strengths and limitations worth understanding.

Advantages and disadvantages of matrix representation

Strengths	Limitations
Turns geometric intuition into concrete calculations that a computer can handle instantly.	The matrix depends on the choice of basis. Different bases give different matrices for the same transformation.
You can compose transformations simply by multiplying their matrices together.	Only works for linear transformations. Non-linear functions (like squaring each coordinate) cannot be captured by a single matrix.
Provides a universal format: any linear transformation, no matter how complicated, fits into the same matrix framework.	For very large dimensions, matrices can become huge and expensive to store and compute with.
Opens the door to powerful tools like determinants, eigenvalues, and inverses.	The matrix entries can sometimes hide the geometric meaning — you lose some visual intuition.

🔑 PERSPECTIVE

Think of the matrix like a GPS app for transformations. The app gives you turn-by-turn arithmetic to reach the answer, which is fast and reliable. But staring at the app's list of directions doesn't give you the same feel for the landscape that looking out the window does. Both views — geometric and algebraic — are valuable.

SECTION 8

Connection to Advanced Theory

The idea of representing a transformation as a matrix is the starting point for many deeper topics in linear algebra and beyond. Here is a preview of where these ideas lead.

From this lesson to more advanced topics

This Lesson	Advanced Extension
Matrix of T using the standard basis	Change of basis: Using a different basis can simplify the matrix (e.g., diagonalization).
Multiplying A · x to get T(x)	Eigenvalues & eigenvectors: Special vectors that T merely stretches without changing direction.
2D and 3D transformations	Infinite dimensions: Functional analysis extends these ideas to spaces of functions, crucial in physics and engineering.
Single transformations	Composition via multiplication: Doing two transformations in a row corresponds to multiplying their matrices, the foundation of computer graphics pipelines.

If you continue studying linear algebra, you'll see that almost every question about linear transformations — finding inverses, measuring how much a transformation stretches area, or decomposing a transformation into simpler pieces — can be answered by working with the matrix. Mastering the matrix representation is like learning to read sheet music: once you can read the notation, an entire world of theory and applications opens up.

SECTION 9

Practice Problems

PROBLEM 1 — CONCEPTUAL

In your own words, explain why knowing T(e₁) and T(e₂) is enough to determine T(x) for every vector x in ℝ². What property of T makes this possible?

PROBLEM 2 — BASIC CALCULATION

A linear transformation T satisfies T(e₁) = (4, −1) and T(e₂) = (2, 5). Write the standard matrix A of T and compute T(3, 1).

PROBLEM 3 — INTERMEDIATE

A linear transformation T: ℝ² → ℝ³ satisfies T(1, 0) = (2, 0, −3) and T(0, 1) = (1, 5, 4). What are the dimensions of the matrix A? Find T(−1, 2).

PROBLEM 4 — APPLIED

A video game rotates every object on screen by 90° counterclockwise and then stretches horizontally by a factor of 2. Write the matrix for the rotation R, the matrix for the stretch S, and find the single combined matrix M = S · R. Use M to transform the point (1, 1).

PROBLEM 5 — CRITICAL THINKING

Suppose T: ℝ² → ℝ² is a linear transformation and you know that T(1, 2) = (5, 3) and T(3, 1) = (7, 4). Can you find the standard matrix of T even though you aren't given T(e₁) and T(e₂) directly? Explain your strategy and find A.

SUMMARY

Lesson Summary

A linear transformation is a function between vector spaces that preserves addition and scalar multiplication. To represent it as a matrix, apply the transformation to each standard basis vector and write the results as columns. The resulting matrix A satisfies T(x) = A · x for every input vector x. Common examples include rotation, reflection, scaling, and shear matrices.

This representation turns geometric actions into arithmetic, making it possible to compute transformations quickly, compose multiple transformations by matrix multiplication, and unlock advanced tools like eigenvalues and change of basis. Remember: the matrix depends on which basis you choose, and it only works for linear transformations.