1. The Standard Introduction

Definition 1.1 – Dot product (numerical)

For two vectors of the same dimension: pair up coordinates, multiply each pair, add.

Worked examples.

Definition 1.2 – Dot product (geometric)

Project onto the line through the origin and the tip of ; then

with a minus sign if the projection points opposite to .

Relative directionSign of
Generally the same direction
Perpendicular (projection is )
Generally opposite

2. Why Order Doesn’t Matter

Theorem 2.1 – Symmetry

The geometric recipe looks asymmetric, but:

  • Equal lengths: projecting onto is a mirror image of projecting onto (reflect across the line of symmetry between them).
  • Scale one vector, say :
    • Projecting onto : the projection length is unchanged, but the vector projected onto is twice as long → doubles.
    • Projecting onto : the projection itself is twice as long → doubles.
    • Either way: . Symmetry is broken in the picture, but not in the result.

The big question remains: why does “multiply matching coordinates and add” have anything to do with projection? → duality.


3. Linear Transformations to the Number Line

Definition 3.1 – Linear map

Takes a 2D vector, returns a number. Visual test of linearity: evenly spaced dots on a line stay evenly spaced on the number line.

Encoding: a 1×2 matrix

and each land on a number, so each column has one entry.

Worked example. , :

Key observation

Multiplying a matrix by a vector is computationally identical to a dot product with that matrix tipped on its side:

matrix⟷2D vector

4. Projection Is a Linear Transformation

Setup

Place a copy of the number line diagonally through the origin. Let be the unit vector whose tip sits at the number 1 on that line.

Define : project any 2D vector onto this diagonal line and read off the number. is linear (passes the evenly-spaced-dots test), so it has a matrix.

Lemma 4.1 – The matrix of projection onto

Where does land? By symmetry (both are unit vectors), projecting onto ‘s line gives the same number as projecting onto the x-axis — which is . Likewise .

Theorem 4.2 – Dot product with a unit vector

= length of the projection of onto the span of (signed).

Theorem 4.3 – Non-unit vectors

Scale by 3: the matrix becomes , which means “project, then multiply by 3”. In general:

This explains why the coordinate formula matches the projection picture.

Worked example. (unit), : , so ‘s shadow on the line through has length 3.


5. Duality

Definition 5.1 – Duality (loose)

A natural but surprising correspondence between two types of mathematical objects.

Theorem 5.2 – Vector ⟷ linear functional

For every linear transformation (however it’s defined), there is a unique vector such that

  • The dual of a vector is the linear transformation it encodes.
  • The dual of a linear map to 1D is a particular vector.

Taking a dot product is a way of turning one of the vectors into a transformation. A vector can be seen as a compact shorthand for a linear map to the number line.

⚠️ Pitfall: The projection line’s outputs are numbers, not 2D vectors — the diagonal copy of the number line is only a visual aid.


Exercises: Compute and interpret the sign; find the vector dual to the map (lesson page).