Linear dependence and independence

• If 𝑉𝑉 is any vector space, then 𝑉𝑉 = Span (𝑉𝑉).

• Clearly, we can �ind smaller sets of vectors which span 𝑉𝑉.
• This lecture we will use the notions of linear independence and linear dependence to �ind the smallest sets
of vectors which span 𝑉𝑉.
• It turns out that there are many "smallest sets" of vectors which span 𝑉𝑉, and that the number of vectors in
these sets is always the same.

This number is the dimension of 𝑽𝑽.

MOTIVATION
1 3 5 1 3 0
We saw that the two sets of vectors ��2� , �5� , � 9 �� and ��2� , �5� , �1�� do not span ℝ3 .
3 7 13 3 7 2
5 1 3 0 1 3
• The problem is that �9 � = 2 �2� + �5� and �1� = 3 �2� − �5�.
13 3 7 2 3 7
• Therefore,
1 3 5 1 3
Span ��2� , �5� , � 9 �� = Span ��2� , �5��
3 7 13 3 7
1 3 0 1 3
and Span ��2� , �5� , �1�� = Span ��2� , �5��.
3 7 2 3 7
• Notice that we can rewrite the two equations above in the following form:
1 3 5 0
2 �2� + �5� − � 9 � = �0� and
3 7 13 0
1 3 0 0
3 �2� − �5� − �1� = �0�
3 7 2 0
This is the key observation about spanning sets.

De�inition:
Suppose that 𝑉𝑉 is a vector space and that 𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑘𝑘 are vectors in 𝑉𝑉.
The set of vectors {𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑘𝑘 } is linearly dependent if
𝑟𝑟1 𝑥𝑥1 + 𝑟𝑟2 𝑥𝑥2 + ⋯ + 𝑟𝑟𝑘𝑘 𝑥𝑥𝑘𝑘 = 0
for some 𝑟𝑟1 , 𝑟𝑟2 , … , 𝑟𝑟𝑘𝑘 ∈ ℝ where at least one of 𝑟𝑟1 , 𝑟𝑟2 , … , 𝑟𝑟𝑘𝑘 is non-zero.
Example:
1 3 5 0 1 3 0 0
2 �2� + �5� − � 9 � = �0� and 3 �2� − �5� − �1� = �0�
3 7 13 0 3 7 2 0
5 1 3 0 1 3
So the two sets of vectors �� 9 � , �2� , �5�� and ��1� , �2� , �5�� are linearly dependent.
13 3 7 2 3 7
Question 1: Suppose that 𝑥𝑥, 𝑦𝑦 ∈ 𝑉𝑉. When are 𝑥𝑥 and 𝑦𝑦 linearly dependent?
Question 2: What do linearly dependent vectors look like in ℝ2 and ℝ3 ?

Example 1:

1 3 0
Let 𝑥𝑥 = �2� 𝑦𝑦 = �2� and 𝑧𝑧 = �4�. Is {𝑥𝑥1 , 𝑥𝑥2 , 𝑥𝑥3 } linearly dependent?
3 1 8
Solution: We must determine whether we can �ind real numbers 𝑟𝑟, 𝑠𝑠, 𝑡𝑡, which are not all zero, such that 𝑟𝑟𝑟𝑟 + 𝑠𝑠𝑠𝑠 +
𝑡𝑡𝑡𝑡 = 0.
To �ind all possible 𝑟𝑟, 𝑠𝑠, 𝑡𝑡 we must solve the augmented matrix equation:
R2 :=𝑅𝑅2 −2𝑅𝑅1
1 3 0 0 𝑅𝑅3 :=𝑅𝑅3 −3𝑅𝑅1 1 3 0 0 𝑅𝑅3 :=𝑅𝑅3 −2𝑅𝑅2 1 3 0 0
�2 2 4 0� ⟶ �0 −4 4 0� ⟶ �0 −4 4 0�
3 1 8 0 0 −8 8 0 0 0 0 0

So this set of equations has a non-zero solution. Therefore, {𝑥𝑥, 𝑦𝑦, 𝑧𝑧} is a linearly dependent set of vectors.
1 3 0 0
To be explicit, 3 �2� − �2� − �4� = �0�.
3 1 8 0
Example 2:

Consider the polynomials 𝑝𝑝(𝑥𝑥) = 1 + 3𝑥𝑥 + 2𝑥𝑥 2 , 𝑞𝑞(𝑥𝑥) = 3 + 𝑥𝑥 + 2𝑥𝑥 2 and 𝑟𝑟(𝑥𝑥) = 2𝑥𝑥 + 𝑥𝑥 2 in ℙ2 . Is {𝑝𝑝(𝑥𝑥), 𝑞𝑞(𝑥𝑥), 𝑟𝑟(𝑥𝑥)}
linearly dependent?

Solution: We must decide whether we can �ind real numbers 𝑟𝑟, 𝑠𝑠, 𝑡𝑡, which are not all zero, such that 𝑟𝑟𝑟𝑟(𝑥𝑥) + 𝑠𝑠𝑠𝑠(𝑥𝑥) +
tr (𝑥𝑥) = 0.
That is:
0 = 𝑟𝑟(1 + 3𝑥𝑥 + 2𝑥𝑥 2 ) + 𝑠𝑠(3 + 𝑥𝑥 + 2𝑥𝑥 2 ) + 𝑡𝑡(2𝑥𝑥 + 𝑥𝑥 2 )
= (𝑟𝑟 + 3𝑠𝑠) + (3𝑟𝑟 + 𝑠𝑠 + 2𝑡𝑡)𝑥𝑥 + (2𝑟𝑟 + 2𝑠𝑠 + 𝑡𝑡)𝑥𝑥 2 .
This corresponds to solving the following system of linear equations.
𝑟𝑟 + 3𝑠𝑠 = 0
3𝑟𝑟 + 𝑠𝑠 + 2𝑡𝑡 = 0
2𝑟𝑟 + 2𝑠𝑠 + 𝑡𝑡 = 0
 1 3 0 1 3 0 𝑅𝑅2 :=𝑅𝑅2 −𝑅𝑅3 1 3 0
𝑅𝑅 :=𝑅𝑅 −3𝑅𝑅
�3 1 2� 𝑅𝑅2 :=𝑅𝑅2 −2𝑅𝑅1 �0 −8 2� ⟶ �0 0 0�
3 3 1
We compute: 2 2 1 0 −4 1 0 −4 1

Hence, {𝑝𝑝(𝑥𝑥), 𝑞𝑞(𝑥𝑥), 𝑟𝑟(𝑥𝑥)} is linearly dependent.

The following sets of vectors are all linearly independent:

• {[1]} is a linearly independent subset of ℝ.
1 0
• �� � , � �� is a linearly independent subset of ℝ2 .
0 1
1 0 0
• ��0� , �1� , �0�� is a linearly independent subset of ℝ3 .
0 0 1
1 0 0 0
0 1 0 0
• �� � , � � , � � , � �� is a linearly independent subset of ℝ4 .
0 0 1 0
0 0 0 1
1 0 0 0
0 1 ⋮ ⋮
• �� � , � � , … , � � , � �� is a linearly independent subset of ℝ𝑚𝑚 .
⋮ ⋮ 1 0
0 0̇ 0 1
• {1} is a linearly independent subset of ℙ0 .
• {1, 𝑥𝑥} is a linearly independent subset of ℙ1 .
• {1, 𝑥𝑥, 𝑥𝑥 2 } is a linearly independent subset of ℙ2 .
• {1, 𝑥𝑥, 𝑥𝑥 2 , … , 𝑥𝑥 𝑛𝑛 } is a linearly independent subset of ℙ𝑛𝑛 .

Example 3:

1 3 5
Let 𝑥𝑥 = �2� 𝑦𝑦 = �2� and 𝑧𝑧 = � 2 �. Is the set {𝑥𝑥1 , 𝑥𝑥2 , 𝑥𝑥3 } linearly independent?
3 9 −1
Solution: We must determine whether or not we can �ind real numbers 𝑟𝑟, 𝑠𝑠, 𝑡𝑡, which are not all zero, such that 𝑟𝑟𝑟𝑟 +
𝑠𝑠𝑠𝑠 + 𝑡𝑡𝑡𝑡 = 0.
Once again, to �ind all possible 𝑟𝑟, 𝑠𝑠, 𝑡𝑡 we must solve the augmented matrix equation:
1 3 5 0 𝑅𝑅2 :=𝑅𝑅2 −2𝑅𝑅1 1 3 5 0 𝑅𝑅2 :=−14𝑅𝑅2 1 3 5 0
�2 2 2 0� 𝑅𝑅2 �0 −4 −8 0� ⟶ �0 1 2 0�
𝑅𝑅3 :=𝑅𝑅3 −3𝑅𝑅1 1
3 9 −1 0 0 0 −16 0 𝑅𝑅3 :=− 𝑅𝑅 0 0 1 0
16 3

Hence, 𝑟𝑟𝑟𝑟 + 𝑠𝑠𝑠𝑠 + 𝑡𝑡𝑡𝑡 = 0 only if 𝑟𝑟 = 𝑠𝑠 = 𝑡𝑡 = 0. Therefore, {𝑥𝑥1 , 𝑥𝑥2 , 𝑥𝑥3 } is a linearly independent subset of ℝ3 .
Example 4:

1 1 1 3
1 2 2 5
Let 𝑥𝑥1 = � � , 𝑥𝑥2 = � � , 𝑥𝑥3 = � � and 𝑥𝑥4 = � �. Is {𝑥𝑥1 , 𝑥𝑥2 , 𝑥𝑥3 , 𝑥𝑥4 } linear dependent or linearly independent?
1 3 1 5
1 4 2 7
Solution: Again, we have to solve the corresponding system of linear equations:
1 1 1 3 1 1 1 3
𝑅𝑅3 =𝑅𝑅3 −2𝑅𝑅2 0 1 1 2 𝑅𝑅4 =𝑅𝑅4 −𝑅𝑅3 0 1 1 2
⟶ � � ⟶ � �
𝑅𝑅4 =𝑅𝑅4 −3𝑅𝑅2 0 0 −2 −2 0 0 −2 −2
0 0 −2 −2 0 0 0 0

Hence, after much work, we see that {𝑥𝑥1 , 𝑥𝑥2 , 𝑥𝑥3 , 𝑥𝑥4 } is linearly dependent.

Example 5:

Let 𝑋𝑋 = {sin 𝑥𝑥, cos 𝑥𝑥} ⊂ 𝔽𝔽. Is 𝑋𝑋 linearly dependent or linearly independent?

Solution: Suppose that 𝑠𝑠 sin 𝑥𝑥 + 𝑡𝑡 cos 𝑥𝑥 = 0.

Notice that this equation holds for all 𝑥𝑥 ∈ ℝ, so

𝑥𝑥 = 0: 𝑠𝑠 ⋅ 0 + 𝑡𝑡 ⋅ 1 = 0
𝜋𝜋
𝑥𝑥 = : 𝑠𝑠 ⋅ 1 + 𝑡𝑡 ⋅ 0 = 0
2
Therefore, we must have 𝑠𝑠 = 0 = 𝑡𝑡. Hence, {sin 𝑥𝑥, cos 𝑥𝑥} is linearly independent. What happens if we tweak this
example by a little bit?

Example 6:
Is {cos 𝑥𝑥, sin 𝑥𝑥, 𝑥𝑥} is linearly independent? If 𝑠𝑠 cos 𝑥𝑥 + 𝑡𝑡sin 𝑥𝑥 + 𝑟𝑟 = 0 then.
𝑥𝑥 = 0: 𝑠𝑠 ⋅ 0 + 𝑡𝑡 ⋅ 1 + 𝑟𝑟 ⋅ 0 =0
𝜋𝜋 𝜋𝜋
𝑥𝑥 = : 𝑠𝑠 ⋅ 1 + 𝑡𝑡 ⋅ 0 + 𝑟𝑟 ⋅ =0
2 2
𝜋𝜋 1 1 𝜋𝜋
𝑥𝑥 = : 𝑠𝑠 ⋅ + 𝑡𝑡 ⋅ + 𝑟𝑟 ⋅ =0
4 √2 √2 4
Therefore, {cos 𝑥𝑥, sin 𝑥𝑥, 𝑥𝑥} is linearly independent.

Example 7:
Show that 𝑋𝑋 = {𝑒𝑒 𝑥𝑥 , 𝑒𝑒 2𝑥𝑥 , 𝑒𝑒 3𝑥𝑥 } is a linearly independent subset of 𝔽𝔽.
Solution: Suppose that 𝑟𝑟𝑒𝑒 𝑥𝑥 + 𝑠𝑠𝑒𝑒 2𝑥𝑥 + 𝑡𝑡𝑒𝑒 3𝑥𝑥 = 0. Then:
𝑥𝑥 = 0 𝑟𝑟 + 𝑠𝑠 + 𝑡𝑡 = 0,
𝑥𝑥 = 1 𝑟𝑟𝑟𝑟 + 𝑠𝑠𝑒𝑒 2 + 𝑡𝑡𝑒𝑒 3 = 0,
𝑥𝑥 = 2 𝑟𝑟𝑒𝑒 2 + 𝑠𝑠𝑒𝑒 4 + 𝑡𝑡𝑒𝑒 6 = 0,
So we must solve the matrix equation:
1
𝑅𝑅2 :=𝑒𝑒−1𝑅𝑅2
1
1 1 1 𝑅𝑅2 := 𝑅𝑅2
𝑒𝑒
1 1 1 1 1 1 1
� 𝑒𝑒 𝑒𝑒 2 𝑒𝑒 3 � ⟶1 �1 𝑒𝑒 𝑒𝑒 2 � 𝑅𝑅3 : = 2 𝑅𝑅3 �0 1 𝑒𝑒 + 1 � 𝑅𝑅3 :=𝑅𝑅3 −𝑅𝑅2 1 1 1
𝑒𝑒 − 1
𝑒𝑒 2 𝑒𝑒 4 𝑒𝑒 6 𝑅𝑅3 :=𝑒𝑒 2 𝑅𝑅3 1 𝑒𝑒 2 𝑒𝑒 4 0 1 𝑒𝑒 2 + 1 ⟶ �0 1 𝑒𝑒 + 1 �
0 0 𝑒𝑒 2 − 𝑒𝑒

Therefore, {𝑒𝑒 𝑥𝑥 , 𝑒𝑒 2𝑥𝑥 , 𝑒𝑒 3𝑥𝑥 } is a set of linearly independent functions in the vector space 𝔽𝔽.
The Basis of a Vector Space:

We now combine the ideas of spanning sets and linear independence.

De�inition: Suppose that 𝑉𝑉 is a vector space. A basis of 𝑉𝑉 is a set of vectors {𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑘𝑘 } in 𝑉𝑉 such that.
• 𝑉𝑉 = Span (𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑘𝑘 ) and
• {𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑘𝑘 } is linearly independent.
Examples:
1 0
• �� � , � � is a basis of ℝ2 .
0 1
1 0 0
• ��0� , �1� , �0�� is a basis of ℝ3 .
0 0 1
1 0 0
0
0 1 ⋮
• �� � , � � , … , � ⋮ � , � �� is a basis of ℝ𝑚𝑚 .
⋮ ⋮ 0
1
0 0̇ 1
• {1, 𝑥𝑥, 𝑥𝑥 2 } is a basis of ℙ2 .
• {1, 𝑥𝑥, 𝑥𝑥 2 , … , 𝑥𝑥 𝑛𝑛 } is a basis of ℙ𝑛𝑛 .
• In general, if 𝑊𝑊 is a vector subspace of 𝑉𝑉 then the challenge is to �ind a basis for 𝑊𝑊.

Summary of Linear Combination and Linear Independence

(1) The expression 𝑐𝑐1 𝐯𝐯1 + 𝑐𝑐2 𝐯𝐯2 + ⋯ + 𝑐𝑐𝑘𝑘 𝐯𝐯𝑘𝑘 is called a linear combination of vectors 𝐯𝐯1 , 𝐯𝐯2 , … , 𝐯𝐯𝑘𝑘 ∈ ℝ𝑛𝑛 , where
𝑐𝑐1 , 𝑐𝑐2 , … , 𝑐𝑐𝑘𝑘 are scalars in ℝ.

(2) A set of vectors {𝐯𝐯1 , 𝐯𝐯2 , … , 𝐯𝐯𝑘𝑘 } is said to be linearly independent if the only scalrs 𝑐𝑐1 , 𝑐𝑐2 , … , 𝑐𝑐𝑘𝑘 satisfying 𝑐𝑐1 𝐯𝐯1 +
𝑐𝑐2 𝐯𝐯2 + ⋯ + 𝑐𝑐𝑘𝑘 𝐯𝐯𝑘𝑘 = 𝟎𝟎 are 𝑐𝑐1 = 𝑐𝑐2 = ⋯ = 𝑐𝑐𝑘𝑘 = 0. We also say that the vectors 𝐯𝐯1 , 𝐯𝐯2 , … , 𝐯𝐯𝑘𝑘 are linearly independent.

(3) If vectors are not linearly independent, they are linearly dependent.

Note:
(1) The set of 𝑛𝑛-dimensional vectors {𝐯𝐯1 , 𝐯𝐯2 , … , 𝐯𝐯𝑘𝑘 } are linearly dependent if 𝑘𝑘 > 𝑛𝑛. (If there are more vectors than
the dimension, then the vectors are linearly dependent.)
 Assignment Problems
2 1 1 1
1. Express the vector 𝐛𝐛 = �13� as a linear combination of the vectors 𝐯𝐯1 = � 5 � , 𝐯𝐯2 = �2� , 𝐯𝐯3 = �4�
6 −1 1 3
1 1 2 2
2. Write the vector � 3 � as a linear combination of the vectors �0� , �−2� , �0�.
−1 0 1 4
3. For what value(s) of 𝑎𝑎 is the following set 𝑆𝑆 linearly dependent?
1 𝑎𝑎 0 1 2
2 0 0 𝑎𝑎 −2
𝑆𝑆 = �� � , � � , � 2 � , � � , � ��
3 −1 𝑎𝑎 1 3
𝑎𝑎 2 7 1 𝑎𝑎3
4. Prove that any set of vectors which contains the zero vector is linearly dependent.
5. Let {𝐯𝐯1 , 𝐯𝐯2 , 𝐯𝐯3 } be a set of nonzero vectors in ℝ𝑚𝑚 such that the dot product 𝐯𝐯𝑖𝑖 ⋅ 𝐯𝐯𝑗𝑗 = 0 when 𝑖𝑖 ≠ 𝑗𝑗. Prove that
the set is linearly independent.

6. 6 Determine whether the following set of vectors is linearly independent or linearly dependent. If the set is
linearly dependent, express one vector in the set as a linear combination of the others.
1 1 −1 −2
0 2 −2 −2
�� � , � � , � � , � ��
−1 3 0 7
0 4 1 11
1 ℎ 1
7. Find the value(s) of ℎ for which the following set of vectors. 𝐯𝐯1 = �0� , 𝐯𝐯2 = � 1 � , 𝐯𝐯3 = � 2ℎ � is linearly
0 −ℎ 3ℎ + 1
independent.
1 1 0
8. Let 𝐯𝐯1 = �2� , 𝐯𝐯2 = �𝑎𝑎� , 𝐯𝐯3 = �4�be vectors in ℝ3 . Determine a condition on the scalars 𝑎𝑎, 𝑏𝑏 so that the set of
0 5 𝑏𝑏
vectors {𝐯𝐯1 , 𝐯𝐯2 , 𝐯𝐯3 } is linearly dependent.
9. Determine conditions on the scalars 𝑎𝑎, 𝑏𝑏 so that the following set 𝑆𝑆 of vectors is linearly dependent.
1 1 0
𝑆𝑆 = {𝐯𝐯1 , 𝐯𝐯2 , 𝐯𝐯3 } where 𝐯𝐯1 = �3� , 𝐯𝐯2 = �𝑎𝑎� , 𝐯𝐯3 = �2�
1 4 𝑏𝑏
1 2
10. Let 𝐴𝐴 be a 3 × 3 matrix and let 𝐯𝐯 = � 2 � and 𝐰𝐰 = �−1�. Suppose that 𝐴𝐴𝐯𝐯 = −𝐯𝐯 and 𝐴𝐴𝐰𝐰 = 2𝐰𝐰.
−1 3
−1
Then �ind the vector 𝐴𝐴5 � 8 �.
−9
11. a) Prove that the column vectors of every 3 × 5 matrix 𝐴𝐴 are linearly dependent.
(b) Prove that the row vectors of every 5 × 3 matrix 𝐵𝐵 are linearly dependent.

12. Suppose 𝑀𝑀 is an 𝑛𝑛 × 𝑛𝑛 upper-triangular matrix. If the diagonal entries of 𝑀𝑀 are all non-zero, then prove that
the column vectors are linearly independent. Does the conclusion hold if we do not assume that 𝑀𝑀 has non-
zero diagonal entries?
13. Suppose that an 𝑛𝑛 × 𝑚𝑚 matrix 𝑀𝑀 is composed of the column vectors 𝐛𝐛1 , ⋯ , 𝐛𝐛𝑚𝑚 . Prove that a vector 𝐯𝐯 ∈ ℝ𝑛𝑛
can be written as a linear combination of the column vectors if and only if there is a vector 𝐱𝐱 which solves the
equation 𝑀𝑀𝐱𝐱 = 𝐯𝐯.