Chapter 4

Vector Spaces

59
4.1 Spaces of Polynomials
4.1.1 Polynomial Vector Space
Definition: Polynomial Space and the Standard Basis of Polynomial Space

The collection of polynomials of degree at most n is denoted Pn . The collection of polynomials given by
{1, x, x2 , ..., xn } is called the monomial basis for Pn .

Definition: Algebra of Polynomials

Let p, q ∈ Pn . If p(x) = an xn + · · · + a1 x + a0 and q(x) = bn xn + · · · + b1 x + b0 and t is a scalar then

polynomial addition is defined as

p(x) + q(x) = (an + bn )xn + · · · + (a1 + b1 )x + (a0 + b0 )

and scalar multiplication is defined as

tp(x) = (tan )xn + · · · + (ta1 )x + (ta0 )

Theorem: Properties of Polynomials

Let p(x), q(x) and r(x) be polynomials in Pn and let s, t ∈ R. Then...

Property Name

p(x) + q(x) is a polynomial of degree at most n Closure under addition

p(x) + q(x) = q(x) + p(x) Commutativity of addition

(p(x) + q(x)) + r(x) = p(x) + (q(x) + r(x)) Associativity of addition

There is a poly. o(x) such that o(x) + s(x) = s(x) for all poly. s(x) Existence of zero

For each s(x) there exists a −s(x) such that s(x) + (−s(x)) = o(x) Existence of inverse

tp(x) is a polynomial of degree at most n Closure under mult.

s(tp(x)) = (st)p(x) Associativity of mult.

(s + t)p(x) = sp(x) + tp(x) Distributivity of scalar add.

t(p(x) + q(x)) = tp(x) + tq(x) Distributivity of poly add.

1 s(x) = s(x) for all s(x) 1 is the scalar identity

60
4.1.2 Linear Combinations, Spans, and Linear Dependendence/Independence
Definition: Linear Combination of Polynomials

A linear combination of polynomials p1 (x),...,pn (x) is given by a sum of the form

t1 p1 (x) + · · · + tn pn (x)

for scalars t1 , ..., tn ∈ R.

Definition: Span of Polynomials

Let B = {p1 (x), ..., pk (x)} be a set of polynomials of degree at most n, Then the span of B is defined as

Span(B) = {t1 p1 (x) + · · · + tk pk (x) | t1 , ..., tk ∈ R}

Example 1: Determine if p(x) = 1 + 2x + 3x2 is in the span of B = {1 + x, 1 + x2 , 1 + x + x2 }. If so, write

p(x) as a linear combination of the polynomials in B.

61
Definition: Linear Dependence and Independence of Polynomials

The set B = {p1 (x), ..., pk (x)} is said to be linearly independent if the only solution to the equation

t1 p1 (x) + · · · + tk pk (x) = 0
is t1 = · · · = tk = 0; otherwise, there is a solution where not all ti are zero, for which B is said to be linearly
dependent.

Example 2: Determine if the set B = {2 − x2 , 3x, −2 + x + x2 } is linearly dependent or independent.

62
4.2 Vector Spaces
4.2.1 Vector Spaces
Definition: Vector Spaces

We refer the reader to the appendix for the requirements of an algebraic space to be called a Vector Space.

Theorem: Examples of Vector Spaces

The following are vector spaces:

1. The space Rn equipped with vector addition and scalar multiplication.

2. The space M (m, n) of all size m × n matrices equipped with matrix addition and scalar multiplication.

3. The space Pn equipped with polynomial addition and scalar multiplication.

4. The space of functions F(a, b) = {f |f : (a, b) ⊆ R → R} equipped with function addition and scalar
multiplication.

Example 1: Consider the space R2 with addition defined to be standard vector addition ⊕ = +, and scalar
multiplication defined by k ⊙ (x, y) = (ky, kx). By finding a counterexample, show that this is not a vector space.

Example 2: Consider the space R with addition defined to be x ⊕ y = x2 + y + 1 and scalar multiplication to
be defined by standard multiplication ⊙ = ·. By finding a counterexample, show that this is not a vector space.

63
 Example 3: Consider R+ = (0, ∞) the space of positive real numbers. Define addition on this space to be
x ⊕ y = xy and scalar multiplication to be s ⊙ x = xs . Prove that this is a vector space.

V1: Closure under addition

V2: Commutativity of addition

V3: Associativity of addition

V4: Existence of zero

V5: Additive inverse

64
V6: Closure under scalar multiplication

V7: Associativity of scalar multiplication

V8: Distributivity of scalar addition

V9: Distributivity of addition

V10: 1 is the scalar identity

Definition: Exponential Space

We will call the space R+ equipped with x ⊕ y = xy and s ⊙ xs the Exponential Space and denote it E.

65
 Theorem: Inverse and Zero Properties of Vector Spaces

Let V be a vector space. Then...

1. 0 ⊙ x = 0 for all x ∈ V

2. (−1) ⊙ x = −x for all x ∈ V

3. t ⊙ 0 = 0 for all t ∈ R

Example 4: Demonstrate the prior theorem holds for E, the exponential space.

Example 5: Let V = {(a, b) | a ∈ R, b ∈ R+ }. Define addition in this space to be (a, b) ⊕ (c, d) = (ad + bc, bd)
and scalar multiplication to be t ⊙ (a, b) = (tabt−1 , bt ). Given this is a vector space, determine the zero vector and
the additive inverse of any vector using the prior theorem.

66
4.2.2 Subspaces
Definition: Subspaces

We refer the reader to the appendix for the definition of a subspace.

Theorem: Subspaces are (“Smaller”) Vector Spaces

A subspace of a vector space is a vector space itself. Consequently, this means that as U ⊆ V you may think
of a vector subspace as a smaller vector space within another vector space.

Example 6: Let U = {p(x) ∈ P3 | p(3) = 0}. Show that U is a subspace of P3 .

Example 7: Prove that the set U = {A ∈ M (2, 2) | a11 + a22 = 0} is a subspace of M (2, 2).

67
4.3 Bases and Dimensions
4.3.1 Linear Combinations, Spans and Bases
Theorem: Spanning Sets as Subspaces

If {v1 , ..., vk } is a set of vectors in a vector space V and S is the set of all possible linear combinations of
these vectors,

S = Span({v1 , ..., vk }) = {(t1 ⊙ v1 ) ⊕ · · · ⊕ (tk ⊙ vk ) | t1 , ..., tk ∈ R}

then S is a subspace of V.

Definition: Spanning Sets of a Vector Space

If S is a subspace of the vector space V consisting of all linear combinations of vectors v1 , ..., vk ∈ V, then S
is called the subspace spanned by B = {v1 , ..., vk }, and we say that the set B spans S. The set B is called
a spanning set for the subspace S. We denote S = Span({v1 , ..., vk }) = Span(B).

Definition: Linear Dependence and Independence

If B = {v1 , ..., vk } is a set of vectors in a vector space V, then B is said to be linearly independent if the
only solution to the equation
(t1 ⊙ v1 ) ⊕ · · · ⊕ (tk ⊙ vk ) = 0
is t1 = · · · = tk = 0; otherwise, there is a non-trivial solution and we say B is said to be linearly dependent.

Theorem: Unique Representation Theorem

Let B = {v1 , ..., vn } be a spanning set for a vector space V. Then every vector in V can be expressed in a
unique way as a linear combination of the vectors of B if and only if the set B is linearly independent.

Definition: Basis of a Vector Space

A set B of vectors in a vector space V is a basis if it is a linearly independent spanning set for V.
     
1 2 0 1 2 5
Example 1: Prove that the set B = , , is not a basis for the subspace Span(B).
−1 1 3 1 1 3
Note: As a reminder, if we don’t specify the operations or vector space, you can assume they are the standard
operations from the standard space. Also, hint... 2A + B − C = O.

68
4.3.2 Determining a Basis of a Subspace
Note: Finding a Basis

Finding a basis of a subspace can be quite difficult. The technique is to first (somehow) determine a spanning
set, then reduce it to or prove that it is, a linearly independent set. Finding the spanning set is the creative
step, while turning a spanning set into a basis is quite procedural. One technique is to associate the span
as the range of some matrix, then discern a basis using our results on the basis of the columnspace of that
matrix.

Example 2: Determine a basis for the subspace S = {p(x) ∈ P2 | p(1) = 0} of P2 . Hint: Every element in this
space can be written as p(x) = (x − 1)(ax + b) for some constants a and b.

Example 3: Determine a basis for Span 1 + x − 2x2 , 2 − x + x2 , 1 − 2x + 3x2 , 1 + 5x + 3x2


.
   
1 2 1 1 1 0 −1 0
RREF
Hint: To speed things up, you are given the fact that  1 −1 −2 5  ∼  0 1 1 0 .
−2 1 3 3 0 0 0 1

69
4.3.3 Dimension
Definition
If B = {v1 , ..., vn } and C = {u1 , ..., uk } are both bases of a vector space V, then k = n. If a vector space
V has a basis with n vectors, then we say that the dimension of V is n and write dim(V) = n. If a vector
space V does not have a basis with finitely many elements, then V is called infinite-dimensional. The
dimension of the trivial vector space V = {0} is defined to be 0.

Example 4: Determine the dimension of the vector space Span(T ) of the previous example.

4.3.4 Extending a Linearly Independent Subset to a Basis

Theorem: Dimension and Linear Independency

Let V be an n-dimensional vector space. Then

1. A set of more than n vectors in V must be linearly dependent.

2. A set of fewer than n vectors cannot span V.

3. A set with n elements of V is a spanning set for V if and only if it is linearly independent.
     
1 1 −2 −1 1 0
Example 5: Let C = , , . Extend C to a basis for M (2, 2). Hint: We’ve
0 1 1 1 1 1
worked with M (2, 2) and know that dim(M (2, 2)) = 4. Thus, you need only find an additional element not in the
span of these three matrices.

70
4.4 Coordinates with Respect to a Basis
4.4.1 Bases
Definition: Coordinate Vectors
Suppose that B = {v1 , ..., vn } is a basis for the vector space V. If x ∈ V with

x = x 1 v1 ⊕ x 2 v2 ⊕ · · · ⊕ x n vn
then the coordinate vector of x with respect to the basis B is
 
x1
 x2 
[x]B =  . 
 
 .. 
xn

Example 1: Consider the bases of R2 (you don’t need to check this)

       
1 1 1 0
B = {⃗v1 , ⃗v2 } = , and E = {⃗e1 , ⃗e2 } = ,
−1 1 0 1
 
3
find [⃗a]B and [⃗a]E where ⃗a = .
−2

Example 2: The collection B = {1, x, 1 + x2 } is a basis of P2 . Find the B-coordinates of p(x) = 2 + x + 3x2 .

71
4.4.2 Change-of-Basis
Theorem: Linearity of Basis Representations

Let B be a basis for a finite dimensional vector space V. Then, for any x, y ∈ V and s, t ∈ R we have

[(s ⊙ x) ⊕ (t ⊙ y)]B = s[x]B + t[y]B

Note: Matrix Representation of Coordinate Representations

This means that the function gB : V → Rn given by gB (v) = [v]B is a linear function. In the event that
V = Rn this means that we should be able to find a representing matrix [gB ].

Note: Development of the Change-of-Basis Matrix

Consider a general vector space V with two bases B and C = {w1 , ..., wn }.

Let x ∈ V and write x be written as a linear combination of the vectors in C,

x = (x1 ⊙ w1 ) ⊕ · · · ⊕ (xn ⊙ wn )
 T
That is, [x]C = x1 x2 · · · xn . Taking B-coordinates gives

[x]B = [(x1 ⊙ w1 ) ⊕ · · · ⊕ (xn ⊙ wn )]B

= x1 [w1 ]B + · · · + xn [wn ]B
 
x1
  . 
= [w1 ]B · · · [wn ]B  .. 
xn
 
= [w1 ]B · · · [wn ]B [x]C

Theorem: Change-of-Basis Matrix

Let B and C = {w1 , ..., wn } both be bases for a vector space V. The matrix
 
P = [w1 ]B · · · [wn ]B
is called the change of coordinates matrix from C-coordinates to B-coordinates and satisfies

[x]B = P [x]C
and is called the change of coordinates equation. Often an emphatic notation PBC is used.

Theorem: Invertibility Reverses the Change-of-Basis, i.e. (PBC )−1 = PCB

Let B and C both be bases for a finite-dimensional vector space V. Let P be the change of coordinates
matrix from C-coordinates to B-coordinates. Then, P is invertible and P −1 is the change of coordinates
matrix from B-coordinates to C-coordinates.

72
 Note: Efficiently Obtaining PBC

Let B = {v1 , v2 , ..., vn }. One will note by prior examples

 that when forming [x]B that you need to solve a
system of linear equations. Hence, when forming PBC = [w1 ]B · · · [wn ]B you need to solve n systems
of linear equations, i.e. t1 ⊙ v1 ⊕ · · · ⊕ tn ⊙ vn = wk for every 1 ≤ k ≤ n. However, the left-hand-
side (i.e. coefficient matrix) is always the same. Hence, you may efficiently solve for PBC by solving the
multi-augmented system
 RREF 
I PBC
 
v1 · · · vn w1 · · · wn ∼

       
1 1 1 0
Example 1: Earlier we considered the bases B = , and E = , for R2 . The
−1 1 0 1
    
E 1 0
change of basis matrix from E to B is PB = , . Demonstrate the prior note by setting up the
0 B 1 B
   
1 0
systems of equations for , and solving for them. Demonstrate the change-of-basis is consistent with
0 B 1 B
T
Example 1 by computing PBE ⃗a where ⃗a = 3 −2 .


73
 Example 2: In this example we introduce another way to obtain a change of basis matrixfrom the standard
basis (specifically) to another basis. Let C = {1, x, x2 } be the standard basis of P2 and let B = 1 − x2 , x, −2 + x2 .
Find the change of coordinates matrix from B to C. Then, use the inverse to find the change of coordinates matrix
from B to C. Do you feel this is conceptually simpler?

74
4.5 General Linear Mappings
4.5.1 General Linearity Conditions
Definition: Linear Mappings

If V and W are vector spaces over R, a function L : V → W is a linear mapping if it satisfies the linearity
properties

Property Name

L(x ⊕V y) = L(x) ⊕W L(y) Additive Linearity

L(t ⊙V x) = t ⊙W L(x) Scalar Linearity

for all x, y ∈ V; t ∈ R; ⊕V and ⊙V are the operations of addition and scalar multiplying respectively on V;
and ⊕W and ⊙W are the operations of addition and scalar multiplying respectively on W. If V = W, then
L may be called a linear operator.

Example 1: Let L : M (2, 2) → P2 be defined by L(A) = a21 + (a12 + a22 )x + a11 x2 . Prove that L is a linear
mapping.

Example 2: Define L : E → R by L(x) = ln(x). Prove that L is a linear function. Note: The function is most
certainly not linear if it is mapping L : R+ → R as usually done in calculus!

75
4.5.2 The General Rank-Nullity Theorem
Definition: The Range and Nullspace/Kernel of a General Linear Mapping

Let V and W be vector spaces over R. The range of a linear mapping L : V → W is defined to be the set

Range(L) = {L(x) ∈ W | x ∈ V}

The nullspace ( or kernel) of L is the set of all vectors in V whose image under L is the zero vector 0W .
We write

Null(L) = {x ∈ V | L(x) = 0W }

Theorem: The Nullspace and Range are Subspaces

Let V and W be vector spaces and let L : V → W be a linear mapping. Then, Null(L) is a subspace of V
and Range(L) is a subspace of W.

Example 3: Consider the mapping L : M (2, 2) → P2 from an earlier example given by L(A) = a21 +
(a12 + a22 )x + a11 x2 . Determine whether 1 + x + x2 ∈ Range(L), and if it is, determine a matrix A such that
L(A) = 1 + x + x2 . Afterwards, determine the nullspace of L.

Theorem: Linear Mappings Fix the Zero Vector

Let V and W be vector spaces and let L : V → W be a linear mapping. Then, L(0V ) = 0W .

Example 4: Demonstrate the prior theorem is true using L : E → R given by L(x) = ln(x) of Example 2.

76
Example 5: Determine a basis for the range and nullspace of the linear mapping L : P1 → R3 defined by
 
0
L(a + bx) =  0 
a − 2b

Definition: The Rank of a General Linear Mapping

Let V and W be vector spaces over R. The rank of a linear mapping L : V → W is the dimension of the
range of L, that is, rank(L) = dim(Range(L)).

Definition: The Nullity of a General Linear Mapping

Let V and W be vector spaces over R. The nullity of a linear mapping L : V → W is the dimension of
the nullspace of L, that is, nullity(L) = dim(Null(L)).

The Rank-Nullity Theorem for a General Linear Mapping

Let V and W be vector spaces over R with dim(V) = n, and let L : V → W be a linear mapping. Then,

rank(L) + nullity(L) = n

Example 6: Confirm the Rank-Nullity Theorem in the previous example.

77
4.6 Matrix of a Linear Mapping
4.6.1 The Matrix of L with Respect to the Basis B
Note: The Representing Matrix of a Mapping in Another Basis

In this subsection we are concerned with the representing matrix [L] of L in another basis B. Specifically, we
want everything stated in the language of B with nothing in the language of the standard basis. That is, for
an input [⃗x]B the output is [L(⃗x)]B and we seek a matrix A such that [L(⃗x)]B = A[⃗x]B , which we will call [L]B .

Let B = {⃗v1 , ..., ⃗vn } be a basis for Rn and let L : Rn → Rn be a linear operator. Then, for any ⃗x ∈ Rn , we
can write ⃗x = b1⃗v1 + · · · + bn⃗vn . Thus by linearity,

L(⃗x) = L(b1⃗v1 + · · · + bn⃗vn ) = b1 L(⃗v1 ) + · · · + bn L(⃗vn )

Representing everything in B coordinates we obtain

[L(⃗x)]B = [b1 L(⃗v1 ) + · · · + bn L(⃗vn )]B

= b1 [L(⃗v1 )]B + · · · + bn [L(⃗vn )]B
 
b1
  . 
= [L(⃗v1 )]B · · · [L(⃗vn )]B  .. 
bn
= [L]B [⃗x]B

Theorem: The Representing Matrix of L in the Basis B

Let V be a vector space. Suppose that B = {v1 , ..., vn } is any basis for V and that L : V → V is a linear
operator. Define the matrix of the linear operator L with respect to the basis B to be the matrix
 
[L]B = [L(v1 )]B · · · [L(vn )]B
where we have for any x ∈ V, [L(x)]B = [L]B [x]B .
   
2 2 1 1
Example 1: Let L : R → R be given by L(x1 , x2 ) = (x2 , x1 ) and let B = , . Determine
−1 1
[L]B .

78
4.6.2 Change of Coordinates and Linear Mappings
Note: Another Way to Obtain [L]B

There is another way to obtain [L]B which is quite simple (in

theory). Specifically, let S represent the standard basis and let
B represent the basis we wish to work in. Let P = PSB be the
change of basis matrix from B to S. Then we may construct

[L]B = PBS [L]PSB ⇒ [L]B = P −1 [L]P

where [L] is just the regular matrix representation of L with
respect to the standard basis (as done in prior chapters). The
logic of this follows the adjacent diagram with linear composition.

Note: Linear Mapping with Different Basis

Using the prior note, you may extend the logic to find the matrix representation of the linear mapping if
the input is in basis B and the output is in basis C. This is simply [L]B S B
C = PC [L]PS , which is very useful.

       
2 1 3  1 1 0 
Example 2: Let L be the linear mapping with [L] = A = −1 2 2 and let B =
   1 , 1 , 1  .
   
−2 3 1 0 1 1
 
Determine [L]B .

79