A matrix moves the plane

Most people meet matrices as spreadsheets with cryptic multiplication rules. Forget that. A 2×2 matrix is a motion of the entire plane — and once you see it move, the multiplication rule becomes the obvious part.

Track just two vectors

A linear transformation is a motion that keeps the grid honest: lines stay lines, the origin stays put, evenly spaced points stay evenly spaced. That discipline has a huge payoff — the whole motion is pinned down by where the two basis vectors land:

That's the entire content of a matrix. The matrix

$A = \begin{pmatrix} 2 & 1 \\ 0 & 1 \end{pmatrix}$

is nothing but the record "î lands at (2, 0), ĵ lands at (1, 1)" — the columns are the landing spots.

Why that determines everything

Any vector is a recipe of basis vectors: v = (3, 2) means 3î + 2ĵ. A linear map respects recipes — so v must land at 3·(where î went) + 2·(where ĵ went). Written out:

$Av = 3\begin{pmatrix}2\\0\end{pmatrix} + 2\begin{pmatrix}1\\1\end{pmatrix} = \begin{pmatrix}8\\2\end{pmatrix}$

That is matrix–vector multiplication. Not a rule to memorize — a consequence of "columns are landing spots" plus linearity.

The gallery of motions

Every 2×2 matrix is some blend of a few pure characters: rotations (spin the plane), scalings (stretch it), shears (slide layers like a deck of cards), reflections (mirror it), and projections (flatten it). Play the gallery in the linear transformation visualizer — type a matrix, scrub the morph slider, and watch the grid carry every vector along.

Why this matters

Rotating a game camera, transforming coordinates between sensors, a layer of a neural network before its activation — all "matrix moves space." Next lesson: a single number that summarizes how violently the space got moved.