[Linear Algebra Series Part 12] Linear Combinations and Span: What Can These Vectors Produce?

What this post covers

This post introduces linear combinations and span.

What it means to mix vectors together
Why span means “everything these vectors can produce”
How this leads to column space
Why this matters in programming, modeling, and representation

Key terms

vector: the basic object we combine
linear combination: a weighted sum of vectors
span: the set of all linear combinations of a given vector set
column space: the span of a matrix’s columns

Core idea

After learning vector addition and scalar multiplication, the next natural question is this:

What can these vectors make together?

That is exactly what linear combinations and span answer.

If v1, v2, ..., vk are vectors and a1, a2, ..., ak are scalars, then

a1v1 + a2v2 + ... + akvk

is a linear combination.

We already know how to scale a vector and add vectors. A linear combination simply does both: scale each vector, then add all the results together.

The span of those vectors is the set of all such results.

This is one of the central ideas in linear algebra because it turns a static list of vectors into a question about expressive power:

what outputs are possible?
what directions are available?
what cannot be produced?

Step-by-step examples

Example 1) The whole plane

Let

v1 = (1, 0), v2 = (0, 1)

Then

a v1 + b v2 = (a, b)

so these two vectors can produce every vector in the plane. Their span is all of R^2.

Example 2) Only one line

Now take

v1 = (1, 2), v2 = (2, 4)

Here v2 = 2v1, so the two vectors are linearly dependent and lie on the same line through the origin. No matter how you combine them, every result stays on that one line. The span is not the whole plane. It is only a line.

This is the first sign that “having more vectors” does not automatically mean “having more expressive power.”

Example 3) A matrix view

Consider

A = [1 0
     0 1
     0 0]

Its columns are [1; 0; 0] and [0; 1; 0]. So every output has the form

x1[1; 0; 0] + x2[0; 1; 0]

That means the third coordinate is always zero. The outputs live in the plane z = 0, not all of R^3.

Example 4) Why this matters for `Ax = b`

Recall that Ax can be viewed as a linear combination of A's columns, where x provides the weights. If the columns of A are a1, a2, ..., an, then

Ax = x1a1 + x2a2 + ... + xnan

So Ax = b is asking whether b belongs to the span of the columns. That span is the column space.

Math notes

When we work over the real numbers, the span of any nonzero vector set is usually infinite, because scalar coefficients can vary continuously.
The zero vector is always in the span, since setting every coefficient to zero produces it.
Span is one of the main ways linear algebra talks about representability.
Later, basis will mean keeping the same span while removing redundancy.

This is why span becomes a bridge to column space, basis, rank, and dimension.

Common mistakes

Thinking span is just a small list of outputs

Usually it is not. It is often an entire line, plane, or higher-dimensional set.

Thinking more vectors always mean a bigger span

Not if the vectors are redundant. Directional diversity matters more than count. By redundant, we mean that one vector can be written as a combination of the others, so it does not add a genuinely new direction.

Treating span as a formula to memorize

It is really a way of asking what a system can express.

Practice or extension

Describe the span of each vector set.

(1, 0), (0, 1)
(1, 2), (2, 4)
(1, 0, 0), (0, 1, 0)
Is (3, 6) in the span of (1, 2) and (2, 4)?
Is (5, -3) in the span of (1, 0) and (0, 1)?

Wrap-up

This post introduced linear combinations and span.

A linear combination mixes vectors by scaling and adding.
Span is the set of everything those mixtures can produce.
The outputs of a matrix are controlled by the span of its columns.
Next, we ask whether a vector set contains redundancy.