(I) Definition and Linear Independence of Matrix Columns

Definition

An indexed set of vectors \(\{\vec{v}_1, \dots, \vec{v}_p\}\) in \(\mathbb{R}^n\) is said to be linearly independent if the vector equation:

\[ x_1\vec{v}_1 + \dots + x_p\vec{v}_p = \vec{0} \]

has only the trivial solution.

The set \(\{\vec{v}_1, \dots, \vec{v}_p\}\) is said to be linearly dependent if there exists weights \(c_1, \dots, c_p\), not all zero, such that:

\[ c_1\vec{v}_1 + \dots + c_p\vec{v}_p = \vec{0} \]

This is called a linear dependence relation among \(\vec{v}_1, \dots, \vec{v}_p\).

If \(A = \begin{bmatrix} \vec{v}_1 & \vec{v}_2 & \dots & \vec{v}_n \end{bmatrix}\) and \(\vec{x} = \begin{bmatrix} x_1 \\ x_2 \\ \vdots \\ x_n \end{bmatrix}\), then:

\[ A\vec{x} = \vec{0} \iff x_1\vec{v}_1 + x_2\vec{v}_2 + \dots + x_n\vec{v}_n = \vec{0} \]

Proposition

The columns of a matrix \(A\) are linearly independent

\(\iff\) The equation \(A\vec{x} = \vec{0}\) has only the trivial solution (no free variable)

\(\iff\) Every column is a pivot column.

Example 1

Let \(\vec{v}_1 = \begin{bmatrix} 1 \\ -1 \\ 4 \end{bmatrix}\), \(\vec{v}_2 = \begin{bmatrix} 3 \\ -5 \\ 7 \end{bmatrix}\), and \(\vec{v}_3 = \begin{bmatrix} -1 \\ 5 \\ 6 \end{bmatrix}\).

  1. Determine if the set \(\{\vec{v}_1, \vec{v}_2, \vec{v}_3\}\) is linearly independent. (This is equivalent to: Determine if the columns of the matrix \(\begin{bmatrix} \vec{v}_1 & \vec{v}_2 & \vec{v}_3 \end{bmatrix}\) are linearly independent.)
  2. If possible, find a linear dependence relation among \(\vec{v}_1\), \(\vec{v}_2\), and \(\vec{v}_3\).

Solution: Construct the coefficient matrix and row reduce:

\[ \begin{bmatrix} 1 & 3 & -1 \\ -1 & -5 & 5 \\ 4 & 7 & 6 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 3 & -1 \\ 0 & -2 & 4 \\ 0 & -5 & 10 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 3 & -1 \\ 0 & 1 & -2 \\ 0 & 0 & 0 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 0 & 5 \\ 0 & 1 & -2 \\ 0 & 0 & 0 \end{bmatrix} \]

Writing the corresponding equations:

\[ \begin{cases} x_1 = -5x_3 \\ x_2 = 2x_3 \\ x_3 \text{ is free} \end{cases} \]

Because there is a free variable, the set \(\{\vec{v}_1, \vec{v}_2, \vec{v}_3\}\) is linearly dependent.

To find a specific linear dependence relation, let \(x_3 = 1\). Then \(x_1 = -5\), \(x_2 = 2\), and \(x_3 = 1\).

A linear dependence relation among \(\vec{v}_1, \vec{v}_2\), and \(\vec{v}_3\) is:

\[ -5\vec{v}_1 + 2\vec{v}_2 + \vec{v}_3 = \vec{0} \]

Note: This relation is not unique. We can find infinitely many linear dependence relations among \(\vec{v}_1\), \(\vec{v}_2\), and \(\vec{v}_3\) by choosing different values for the free variable.

Example 2

Find the value(s) of \(h\) for which the vectors are linearly dependent:

\[ \begin{bmatrix} 2 \\ -4 \\ 1 \end{bmatrix}, \begin{bmatrix} -6 \\ 7 \\ -3 \end{bmatrix}, \begin{bmatrix} 8 \\ h \\ 4 \end{bmatrix} \]

Solution: Construct the coefficient matrix and row reduce to find the pivot columns:

\[ \begin{bmatrix} 2 & -6 & 8 \\ -4 & 7 & h \\ 1 & -3 & 4 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & -3 & 4 \\ 0 & -5 & h+16 \\ 0 & 0 & 0 \end{bmatrix} \]

Notice that \(x_3\) is a free variable. The equation \(x_1\vec{v}_1 + x_2\vec{v}_2 + x_3\vec{v}_3 = \vec{0}\) has a free variable and hence a nontrivial solution no matter what the value of \(h\) is.

So the vectors are linearly dependent for all values of \(h\).

(II) Sets of Vectors

Case 1: Sets of One Vector

The set \(\{\vec{v}_1\}\) is linearly independent if and only if \(x_1\vec{v}_1 = \vec{0}\) has only the trivial solution \(x_1 = 0\). This happens if and only if \(\vec{v}_1 \neq \vec{0}\).

Conclusion:
  1. \(\{\vec{0}\}\) is always linearly dependent.
  2. A set of one vector \(\{\vec{v}_1\}\) is linearly independent if and only if \(\vec{v}_1 \neq \vec{0}\).

Case 2: Sets of Two Vectors \(\{\vec{v}_1, \vec{v}_2\}\)

The set is linearly dependent if and only if \(x_1\vec{v}_1 + x_2\vec{v}_2 = \vec{0}\) has a nontrivial solution.

Conclusion:
  1. \(\{\vec{v}_1, \vec{v}_2\}\) is linearly dependent if at least one of the vectors is a multiple of the other.
  2. \(\{\vec{v}_1, \vec{v}_2\}\) is linearly independent if and only if neither of the vectors is a multiple of the other.

Note: We can always determine by inspection when a set of 2 vectors is linearly dependent.

Case 3: Sets of Two or More Vectors \(S = \{\vec{v}_1, \dots, \vec{v}_p\}\)

Theorem 7 (Characterization of Linearly Dependent Sets)

An indexed set \(S = \{\vec{v}_1, \dots, \vec{v}_p\}\) of two or more vectors is linearly dependent if and only if at least one of the vectors in \(S\) is a linear combination of the others.

In fact, if \(S\) is linearly dependent and \(\vec{v}_1 \neq \vec{0}\), then some \(\vec{v}_j\) (with \(j > 1\)) is a linear combination of the preceding vectors \(\vec{v}_1, \dots, \vec{v}_{j-1}\).

(Proof in textbook)

Example 3 (Practice Problem)

Let \(\vec{u} = \begin{bmatrix} 3 \\ 2 \\ -4 \end{bmatrix}\), \(\vec{v} = \begin{bmatrix} -6 \\ 1 \\ 7 \end{bmatrix}\), \(\vec{w} = \begin{bmatrix} 0 \\ -5 \\ 2 \end{bmatrix}\), and \(\vec{z} = \begin{bmatrix} 3 \\ 7 \\ -5 \end{bmatrix}\).

(a) Are the sets \(\{\vec{u}, \vec{v}\}\), \(\{\vec{u}, \vec{w}\}\), \(\{\vec{u}, \vec{z}\}\), \(\{\vec{v}, \vec{w}\}\), \(\{\vec{v}, \vec{z}\}\), \(\{\vec{w}, \vec{z}\}\) each linearly independent?

Yes. In each case, neither vector is a scalar multiple of the other.

(b) Does the answer to Part (a) imply that \(\{\vec{u}, \vec{v}, \vec{w}, \vec{z}\}\) is linearly independent?

The answer to Part (a) says nothing about the linear independence of the set of all four vectors.

(c) Is \(\vec{w}\) a linear combination of \(\vec{u}\), \(\vec{v}\), and \(\vec{z}\)?

Form the augmented matrix \(\begin{bmatrix} \vec{u} & \vec{v} & \vec{z} & | & \vec{w} \end{bmatrix}\):

\[ \begin{bmatrix} 3 & -6 & 3 & 0 \\ 2 & 1 & 7 & -5 \\ -4 & 7 & -5 & 2 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & -2 & 1 & 0 \\ 0 & 1 & 1 & -1 \\ 0 & 0 & 0 & 1 \end{bmatrix} \]

The system is inconsistent. Therefore, \(\vec{w}\) is not a linear combination of \(\vec{u}\), \(\vec{v}\), and \(\vec{z}\).

Note: When testing for linear independence, it's usually a poor idea to check if one selected vector is a linear combination of the others.

(d) Is \(\{\vec{u}, \vec{v}, \vec{w}, \vec{z}\}\) linearly independent?

Form the coefficient matrix \(\begin{bmatrix} \vec{u} & \vec{v} & \vec{w} & \vec{z} \end{bmatrix}\):

\[ \begin{bmatrix} 3 & -6 & 0 & 3 \\ 2 & 1 & -5 & 7 \\ -4 & 7 & 2 & -5 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & -2 & 0 & 1 \\ 0 & 1 & -1 & 1 \\ 0 & -1 & 2 & -1 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & -2 & 0 & 1 \\ 0 & 1 & -1 & 1 \\ 0 & 0 & 1 & 0 \end{bmatrix} \]

\(x_4\) is a free variable. Therefore, \(\{\vec{u}, \vec{v}, \vec{w}, \vec{z}\}\) is linearly dependent. (Or use Theorem 8 below).

Warning: Theorem 7 doesn't say that every vector in a linearly dependent set is a linear combination of the preceding vectors. A vector in a linearly dependent set may fail to be a linear combination of the other vectors.

The next two theorems describe special cases in which the linear dependence of the set is automatic.

Theorem 8

If a set contains more vectors than there are entries in each vector, then the set is linearly dependent.

Proof:

Let \(A = \begin{bmatrix} \vec{v}_1 & \dots & \vec{v}_p \end{bmatrix}\).

If \(A\vec{x} = \vec{0}\) has more unknowns than the number of equations, then \(A\) has a larger number of columns than the number of rows.

There must be free variables, meaning nontrivial solutions exist.

\(\implies \{\vec{v}_1, \dots, \vec{v}_p\}\) is linearly dependent.

Theorem 9

Any set containing \(\vec{0}\) is linearly dependent.

Proof:

Suppose \(\vec{v}_1 = \vec{0}\).

Then \(1\cdot\vec{v}_1 + 0\cdot\vec{v}_2 + \dots + 0\cdot\vec{v}_p = \vec{0}\) is a valid linear dependence relation (since at least one weight, the first, is non-zero).

Example 4

Determine by inspection if the given set is linearly dependent.

  1. \(\left\{ \begin{bmatrix} 3 \\ 2 \\ 0 \end{bmatrix}, \begin{bmatrix} 6 \\ 0 \\ 4 \end{bmatrix} \right\}\)
    Answer: Linearly independent. (Neither is a scalar multiple of the other).
  2. \(\left\{ \begin{bmatrix} 1 \\ 7 \\ 6 \end{bmatrix}, \begin{bmatrix} 2 \\ 0 \\ 9 \end{bmatrix}, \begin{bmatrix} 3 \\ 1 \\ 0 \end{bmatrix}, \begin{bmatrix} 4 \\ 1 \\ 8 \end{bmatrix} \right\}\)
    Answer: Linearly dependent by Theorem 8 (4 vectors in \(\mathbb{R}^3\)).
  3. \(\left\{ \begin{bmatrix} 1 \\ 2 \\ 0 \\ 1 \end{bmatrix}, \begin{bmatrix} -1 \\ 5 \\ 3 \\ 2 \end{bmatrix}, \begin{bmatrix} 0 \\ 0 \\ 0 \\ 0 \end{bmatrix} \right\}\)
    Answer: Linearly dependent by Theorem 9 (contains the zero vector).
Example 5

(1) How many pivot columns must a \(5 \times 4\) matrix have if its columns are linearly independent? Why?

Solution: All 4 columns of the \(5 \times 4\) matrix must be pivot columns. Otherwise, \(A\vec{x} = \vec{0}\) would have a free variable, in which case the columns of \(A\) would be linearly dependent.

(2) How many pivot columns must a \(4 \times 5\) matrix \(A\) have if its columns span \(\mathbb{R}^4\)? Why?

Solution: If the matrix spans \(\mathbb{R}^4\), then \(A\) has a pivot in each row by Theorem 4. Since each pivot position is in a different column, \(A\) has 4 pivot columns.