Chapter 3 · Structure

Vector Spaces and Linear Transformations

One question

What does a matrix keep, collapse, and reach?

Before the code

Four ideas to keep in view

Follow an input through the matrix and ask what survives at the output.

01Reachability
b∈Col⁡(A)b \in \operatorname{Col}(A)

A solution exists exactly when b is reachable from the columns of A.

02Uniqueness
Null⁡(A)={0}\operatorname{Null}(A) = \{0\}

If a solution exists, it is unique exactly when no nonzero input direction disappears.

03Dimension
rank⁡(A)+nullity⁡(A)=n\operatorname{rank}(A) + \operatorname{nullity}(A) = n

For A∈Rm×nA \in \mathbb R^{m\times n}, the number of pivots plus the number of free variables equals n, the number of input coordinates.

04Linear map
A(cx+y)=cAx+AyA(cx+y)=cAx+Ay

A matrix preserves vector addition and scalar multiplication.

Structure bridge

Different inputs can reach the same output

For A=[[1,2],[2,4]]A = [[1,2],[2,4]], the nonzero direction z=(−2,1)Tz = (-2,1)^{\mathsf T} disappears. Adding z to an input cannot change the output. Each row in the display below lists one vector’s coordinates.

Inputs · x and x + z
30
11
  1. 01z = (−2, 1)ᵀa null direction
  2. 02Az = 0the direction disappears
  3. 03A(x + z) = Axthe output stays fixed
Outputs · same b
36
36

Chapter 3Null space controls uniqueness; column space controls existence.

Practice sequence

Choose a lab

  1. 3.1Finding the Null SpaceTurn one free variable into a direction that the matrix sends to zero.eliminate → split columns → build a basisOpen lab
  2. 3.2Interpretable Linear TransformationsFollow one vector through rotation, projection, and reflection.rotate → project → reflectOpen lab
  3. 3.3Solving Non-invertible Linear SystemsFollow Lecture 3’s 3 × 4 matrix from elimination to a consistency condition and the complete two-parameter solution.eliminate → test existence → find every solutionOpen lab