Matrix Transformations of Vectors

SECTION 1

Historical Context & Motivation

Long before computers could render 3-D video game worlds or animate movie characters, mathematicians were searching for a compact, powerful way to describe geometric operations like rotation, reflection, and scaling. The answer turned out to be matrices — rectangular arrays of numbers that can act on vectors (arrows with both direction and magnitude) to transform them in predictable ways. The idea that a single multiplication could encode an entire geometric transformation was a breakthrough that changed mathematics, physics, and eventually computer science.

1850s

Cayley Formalizes Matrices

British mathematician Arthur Cayley publishes the first systematic theory of matrices, defining matrix multiplication and showing that matrices form an algebraic system.

1907

Matrices Enter Physics

Hermann Minkowski uses matrix-like notation to express Einstein's special relativity, demonstrating that physical transformations of space and time can be written as matrix operations on four-dimensional vectors.

1950s

Computer Graphics Pioneers

Early computer scientists realize that every on-screen rotation, scaling, or translation of a shape can be computed by multiplying vertex coordinates (vectors) by transformation matrices, launching the field of computer graphics.

Today

Everyday Applications

Matrix-vector multiplication drives GPS satellite corrections, robotic arm movements, image filters on your phone, and the rendering of every frame in modern video games and animated films.

The central question this lesson answers is straightforward: How do we multiply a matrix by a vector, and what does the result mean geometrically? Once you can answer that, you hold the key to understanding every geometric transformation that a computer — or a mathematician — can describe with linear algebra.

SECTION 2

Core Principles & Definitions

Before we multiply anything, let's pin down the vocabulary. A vector in two dimensions is an ordered pair like [x, y], which we write as a column (a matrix with one column and two rows). A matrix is a rectangular array of numbers arranged in rows and columns. When we multiply a 2 × 2 matrix by a 2 × 1 column vector, we get a new 2 × 1 column vector — a transformed version of the original.

Column Vector

A vector written as a single-column matrix. In 2-D, it has 2 rows and 1 column. Example: the point (3, 5) becomes the column vector with entries 3 and 5.

Transformation Matrix

A square matrix (often 2 × 2) that encodes a geometric operation — rotation, reflection, scaling, or shearing — applied to every vector it multiplies.

Dimension Compatibility

Matrix A (m × n) can multiply vector v (n × 1) only if the number of columns of A equals the number of rows of v. The result is an m × 1 vector.

Dot-Product Rule

Each entry of the resulting vector is the dot product of the corresponding row of the matrix with the entire column vector.

Linearity

Matrix transformations are linear: they preserve addition of vectors and scalar multiplication. This means straight lines stay straight and the origin stays fixed.

✦ KEY TAKEAWAY

Think of a matrix as a machine and a vector as the raw material you feed in. The machine takes the input vector, stretches it, rotates it, flips it — whatever the matrix's numbers dictate — and spits out a brand-new vector. Different matrices are different machines; the same input vector can come out looking completely different depending on which machine processes it.

SECTION 3

Visual Explanation — Seeing the Transformation

The diagram below shows a 2-D coordinate grid with an original vector v (in cyan) and its image Av (in pink) after multiplication by a specific 2 × 2 matrix A. Notice how the transformation changes both the direction and the length of the vector while keeping the origin fixed — the hallmark of a linear transformation.

The cyan arrow is the original vector v = (2, 1). After multiplication by matrix A, the transformed vector Av = (4, 3) (pink) is both longer and rotated. The dashed amber arc highlights the change in angle.

In the diagram above, the matrix A combines two effects at once: it stretches the vector (making it longer) and rotates it (changing its angle). That is the power of matrix multiplication — a single compact operation can encode a combination of geometric moves. Every point in a shape could be treated as a vector, multiplied by the same matrix, and the entire shape transforms uniformly.

SECTION 4

Mathematical Framework

Let's formalize the multiplication rule. Given a 2 × 2 matrix and a 2 × 1 column vector, the product is computed row by row using dot products. Each row of the matrix pairs with the column vector, and you multiply corresponding entries and add. The same pattern extends to 3 × 3 matrices with 3 × 1 vectors, and beyond.

MATRIX–VECTOR MULTIPLICATION (2 × 2)

[ a b ] × [ x ] = [ a·x + b·y ] [ c d ] [ y ] [ c·x + d·y ]

The matrix has entries a, b, c, d. The input vector has components x and y. The first entry of the result is the dot product of row 1 with the vector; the second entry is the dot product of row 2 with the vector.

ROTATION MATRIX (angle θ)

R(θ) = [ cos θ −sin θ ] [ sin θ cos θ ]

This matrix rotates any 2-D vector counterclockwise by angle θ about the origin. Multiplying R(θ) by a column vector produces the rotated vector.

SCALING MATRIX

S = [ s₁ 0 ] [ 0 s₂ ]

This diagonal matrix scales the x-component by s₁ and the y-component by s₂. If s₁ = s₂, the scaling is uniform (the vector stretches or shrinks without changing direction).

REFLECTION ACROSS THE x-AXIS

M = [ 1 0 ] [ 0 −1 ]

This matrix flips the y-component's sign, reflecting any vector across the x-axis. Multiplying by M sends the point (x, y) to (x, −y).

⚠️ Dimension Check

A matrix of size m × n can only multiply a vector of size n × 1. The inner dimensions must match. The result is a vector of size m × 1. If you try to multiply a 2 × 2 matrix by a 3 × 1 vector, the operation is undefined.

SECTION 5

Common Transformation Types

Different matrices produce different geometric effects. The diagram below catalogues four fundamental 2-D transformations — rotation, scaling, reflection, and shearing — each applied to the same square shape. Understanding these four building blocks lets you decode any 2 × 2 transformation matrix you encounter.

Four fundamental transformations applied to a unit square. The dashed outline is the original shape; the solid colored outline is the transformed shape. Each panel shows the corresponding 2 × 2 matrix.

Summary of common 2-D transformation matrices

Transformation	Matrix Form	Geometric Effect
Rotation by θ	`[cos θ −sin θ; sin θ cos θ]`	Turns every vector counterclockwise by θ about the origin; preserves length.
Scaling	`[s₁ 0; 0 s₂]`	Stretches or compresses along x by s₁ and along y by s₂.
Reflection (x-axis)	`[1 0; 0 −1]`	Flips the y-component; mirrors across the x-axis.
Reflection (y-axis)	`[−1 0; 0 1]`	Flips the x-component; mirrors across the y-axis.
Horizontal shear	`[1 k; 0 1]`	Slides points horizontally in proportion to their y-value; turns rectangles into parallelograms.

SECTION 6

Worked Example

Let's walk through a complete matrix-vector multiplication and then interpret the result geometrically. Suppose we want to rotate the vector v = (3, 1) by 90° counterclockwise.

Rotating v = (3, 1) by 90° Counterclockwise

Step 1 — Identify the Transformation MatrixFor a 90° counterclockwise rotation, cos 90° = 0 and sin 90° = 1. The rotation matrix is R(90°) = [0 −1; 1 0].

R = [0 −1; 1 0]

Step 2 — Write the Vector as a ColumnThe vector v = (3, 1) is written as a 2 × 1 column matrix: [3; 1]. This ensures the inner dimensions match: the matrix is 2 × 2 and the vector is 2 × 1.

v = [3; 1]

Step 3 — Compute Row 1 Dot ProductTake row 1 of R, which is (0, −1), and dot it with v = (3, 1). That gives: (0)(3) + (−1)(1) = 0 + (−1) = −1. This is the x-component of the result.

First entry = −1

Step 4 — Compute Row 2 Dot ProductTake row 2 of R, which is (1, 0), and dot it with v = (3, 1). That gives: (1)(3) + (0)(1) = 3 + 0 = 3. This is the y-component of the result.

Second entry = 3

Step 5 — State the Result and InterpretThe transformed vector is Rv = (−1, 3). If you plot both vectors, you'll see that v = (3, 1) pointed mostly to the right, and the result (−1, 3) points mostly upward — exactly a 90° counterclockwise rotation. You can also verify the length is preserved: |v| = √(9 + 1) = √10, and |Rv| = √(1 + 9) = √10.

Rv = (−1, 3)

SECTION 7

Strengths & Limitations

Matrix transformations are incredibly useful, but they do have boundaries. Understanding both their strengths and their limitations will help you know when to use them — and when a different tool is needed.

Matrix transformations: strengths vs. limitations

Strengths	Limitations
Compact: a single 2 × 2 matrix encodes an entire geometric transformation.	Cannot represent translations (sliding a shape without rotating) with a standard 2 × 2 matrix; you need 3 × 3 augmented matrices for that.
Composable: multiplying two matrices gives a matrix for the combined transformation.	Order matters — matrix multiplication is not commutative, so rotating then scaling differs from scaling then rotating.
Efficient: computers can multiply millions of vectors by a matrix almost instantly using optimized hardware (GPUs).	Limited to linear transformations; non-linear warps (like a fish-eye lens effect) require more advanced techniques.
Universal: works in 2-D, 3-D, and higher dimensions with the same dot-product rule.	Larger matrices (e.g., 4 × 4 for 3-D graphics) can be harder to visualize and debug by hand.

✦ KEY TAKEAWAY

Matrix transformations are like a recipe card for geometry. One card can say "rotate 45° and double the size," and you can apply that card to every point in a shape just by doing one multiplication each time. The limitation is that the recipe card can only handle operations that keep the origin pinned and lines straight — if you need to slide everything two units to the right, you need a slightly bigger card (an augmented matrix).

SECTION 8

Connection to Advanced Theory

The 2 × 2 matrix-vector multiplication you've learned is the gateway to a much larger world. In a college linear algebra course, you'll study eigenvalues and eigenvectors — special vectors that don't change direction under a given matrix, only get scaled. You'll also encounter determinants, which tell you by what factor a matrix scales area. These ideas extend naturally from the multiplication rule you've practiced here.

From high school to college linear algebra

This Lesson (CCSS N-VM.11)	College Linear Algebra
2 × 2 matrices act on 2-D vectors	n × n matrices act on n-dimensional vectors; same dot-product rule
Recognize rotation, scaling, reflection by inspecting matrix entries	Eigenvalue decomposition reveals the fundamental stretches and rotations hidden in any matrix
Compose two transformations by multiplying matrices	Matrix groups and abstract algebra formalize all possible compositions
Dimension compatibility: inner dimensions must match	Rank, null space, and the dimension theorem explain when and why systems have solutions

If you continue in STEM fields, matrix-vector multiplication will become one of your most-used tools. In physics it describes quantum states, in data science it powers machine learning algorithms, and in engineering it models everything from stress in beams to signal processing. The core operation — row-by-column dot products — never changes, no matter how advanced the application.

SECTION 9

Practice Problems

PROBLEM 1 — CONCEPTUAL

A 2 × 2 matrix A is multiplied by a 2 × 1 column vector v. What are the dimensions of the resulting vector? Explain why those dimensions make sense in terms of the dot-product rule.

PROBLEM 2 — BASIC CALCULATION

Compute the product: [2 3; 1 4] × [5; −1].

PROBLEM 3 — INTERMEDIATE

The matrix [0 −1; 1 0] represents a 90° counterclockwise rotation. Verify this by applying it to the vector v = (1, 0) and then to w = (0, 1). Do the results match what you'd expect from a 90° rotation?

PROBLEM 4 — APPLIED

A video game developer needs to reflect a character's sprite across the y-axis so it faces the opposite direction. The character's nose is at position (4, 3) relative to the character's center. Write the appropriate 2 × 2 reflection matrix and compute the new nose position.

PROBLEM 5 — CRITICAL THINKING

Show that applying the 90° counterclockwise rotation matrix R = [0 −1; 1 0] four times in a row returns any vector to its original position. (Hint: compute R × R to get R², then keep going, or reason geometrically.)

SUMMARY

Lesson Summary

In this lesson you learned that a vector can be written as a column matrix and multiplied by a transformation matrix using the dot-product rule: each entry of the result equals a row of the matrix dotted with the entire vector. The dimension compatibility requirement states that an m × n matrix can only act on an n × 1 vector, producing an m × 1 vector.

You explored four fundamental types of 2-D transformation — rotation, scaling, reflection, and shearing — each encoded by a specific 2 × 2 matrix. These transformations are linear, meaning they keep the origin fixed and preserve straight lines. By composing (multiplying) matrices, you can chain transformations together, and this single operation — matrix-vector multiplication — forms the foundation for everything from computer graphics to advanced physics.