# Span and linear independence

** Published:**

*An extremely important concept linear algebra is that of linear independence. In this blog post we present the definition for the span of a set of vectors. Then, we use this definition to discuss the definition for linear independence. Finally, we discuss some intuition into this fundamental idea.*

## Introduction

An extremely important concept in the study of vector spaces is that of *linear independence*. At a high level, a set of vectors are said to be **linearly independent** if you cannot form any vector in the set using any combination of the other vectors in the set. If a set of vectors does not have this quality – that is, a vector in the set can be formed from some combination of others – then the set is said to be **linearly dependent**.

In this post, we will present a more foundatioanl concept, the *span* of a set of vectors, and then move on to the definition for linear independence. Finally, we will discuss a high-level intuition for why the concept of linearly independence is so important.

## Span

Given a set of vectors, the **span** of the set of vectors are all of the vectors that can be “constructed” by taking linear combinations of vectors in that set. More rigorously,

**Definition 1 (span):** Given a vector space, $(\mathcal{V}, \mathcal{F})$ and a set of vectors $S := \boldsymbol{x}_1, \boldsymbol{x}_2, \dots, \boldsymbol{x}_n \in \mathcal{V}$, the **span** of $S$, denoted $\text{Span}(S)$ is the set of all vectors that can be formed by taking a linear combination of vectors in $S$. That is,

Intuitively, you can think of $S$ as a set of “building blocks” and the $\text{Span}(S)$ as the set of all vectors that can be “constructed” from the building blocks in $S$. To illustrate this point, we show in the figure below two vectors, $\boldsymbol{x}_1$ and $\boldsymbol{x}_2$ (left), and two examples of vectors in their span (right):

Note, we can see that in this example that we could construct *ANY* two dimensional vector from $\boldsymbol{x}_1$ and $\boldsymbol{x_2}$. Thus, the span of these two vectors is all of $\mathbb{R}^2$! This is not always the case. In the figure below, we show an example with a different span:

This time, $\boldsymbol{x}_1$ and $\boldsymbol{x_2}$ don’t span all of $\mathbb{R}^2$, but rather, only the line on which $\boldsymbol{x}_1$ and $\boldsymbol{x_2}$ lie.

## Linear independence

Given a vector space, $(\mathcal{V}, \mathcal{F})$, and a set of vectors $S := \boldsymbol{x}_1, \boldsymbol{x}_2, \dots, \boldsymbol{x}_n \in \mathcal{V}$, the vectors are said to be **linearly independent** if each vector lies outside the span of the remaining vectors. More rigorously,

**Definition 2 (linear independence):** Given a vector space, $(\mathcal{V}, \mathcal{F})$ and a set of vectors $S := \boldsymbol{x}_1, \boldsymbol{x}_2, \dots, \boldsymbol{x}_n \in \mathcal{V}$, $S$ is called **linearly independent** if for each vector $\boldsymbol{x_i} \in S$, it holds that $\boldsymbol{x}_i \notin \text{Span}(S \setminus {\boldsymbol{x}_i})$.

Said differently, a set of vectors are linearly independent if you cannot form any of the vectors in the set using a linear combination of any of the other vectors. Below we demonstrate a set of linearly independent vectors (left) and a set of linearly dependent vectors (right):

Why is the set on the right linearly dependent? As you can see below, we can use any of the two vectors to construct the third:

## Intuition

There are two ways I think about linear independence: in terms of information content and in terms of intrinsic dimensionality. Let me explain.

First, if a set of vectors is linearly dependent, then in a sense there is “reduntant information” within the vectors. What do we mean by redundant? By removing a vector from a linearly dependent set of vectors, the span of the set of vectors will remain the same! On the other hand, for a linearly independent set of vectors, each vector is vital for defining the span of the set’s vectors. If you remove even one vector, the span of the vectors will change (in fact, it will become smaller)!

At a more geometric level of thinking, a set of $n$ linearly independent vectors $S := { \boldsymbol{x}_1, \dots, \boldsymbol{x}_n }$ spans a space with an intrinsic dimensionality of $n$ because in order to specify any vector $\boldsymbol{v}$ in the span of these vectors, one must specify the coefficients $c_1, \dots, c_n$ to construct $\boldsymbol{v}$ from the vectors in $S$. That is,

\[\boldsymbol{v} = c_1\boldsymbol{x}_1 + \dots + c_n\boldsymbol{x}_n\]However, if $S$ is linearly dependent, then we can throw away “redundant” vectors in $S$. In fact, we see that the intrinsic dimensionality of a linearly dependent set $S$ is the maximum sized subset of $S$ that is linearly independent!