MENU Intensive Linear Algebra

c 2013 John M. Alongi. All Rights Reserved.

September 22, 2013

Contents
1 Vectors
1.1 Geometric Vectors . . . . . . . .
1.2 Rectangular Coordinate Systems
1.3 Coordinate Vectors . . . . . . . .
1.4 The Vector Sum . . . . . . . . .
1.5 The Scalar-Vector Product . . .
1.6 Lines . . . . . . . . . . . . . . . .
1.7 Planes . . . . . . . . . . . . . . .
1.8 Linear Combinations . . . . . . .
2 The
2.1
2.2
2.3
2.4
2.5
2.6

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

1
1
6
13
18
23
29
35
40

Inner Product
The Inner Product: Motivation . . . . . . . . . . . . .
The Inner Product: Definition and Properties . . . . .
Norm and Distance . . . . . . . . . . . . . . . . . . . .
Orthogonality . . . . . . . . . . . . . . . . . . . . . . .
Orthogonal Decomposition and Orthogonal Projection
The Cauchy-Schwarz Inequality . . . . . . . . . . . . .

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

47
47
49
55
61
63
67

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

3 Systems of Linear Equations

3.1 Systems of Linear Equations . . . . . . . . . .
3.2 Matrices, Row Reduction and Echelon Forms
3.3 Solving Systems of Linear Equations . . . . .
3.4 Span . . . . . . . . . . . . . . . . . . . . . . .
3.5 The Matrix-Vector Product . . . . . . . . . .
3.6 Solution Sets of Linear Systems . . . . . . . .
3.7 Linear Independence . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

75
75
84
92
105
109
114
118

4 Linear Transformations
4.1 Linear Transformations: Definition and Properties .
4.2 The Standard Matrix of a Linear Transformation . .
4.3 The Algebra of Linear Transformations and Matrices
4.4 Surjectivity and Injectivity . . . . . . . . . . . . . .
4.5 Invertibility . . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

127
127
137
140
152
160

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

ii
5 Linear Spaces
5.1 Linear Spaces . . . . . . . . . . . .
5.2 Subspaces . . . . . . . . . . . . . .
5.3 Span . . . . . . . . . . . . . . . . .
5.4 Linear Independence . . . . . . . .
5.5 Bases and Dimension . . . . . . . .
5.6 Linear Transformations . . . . . .
5.7 Images, Kernels, and Isomorphisms
5.8 The Rank-Nullity Theorem . . . .
5.9 Coordinates . . . . . . . . . . . . .
5.10 Change of Basis . . . . . . . . . . .

CONTENTS

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

173
173
179
185
188
192
200
204
208
210
217

A Foundations
223
A.1 Mathematical Induction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
A.2 The Real Number System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
A.3 The Complex Number System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
B Table of Equivalent Statements
C Answers to Selected Exercises
C.1 Chapter 1: Vectors . . . . . . . . . . . .
C.2 Chapter 2: Inner Product . . . . . . . .
C.3 Chapter 3: Systems of Linear Equations
C.4 Chapter 4: Linear Transformations . . .
C.5 Chapter 5: Linear Spaces . . . . . . . .
C.6 Appendix A: Foundations . . . . . . . .

245
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

247
247
250
252
257
261
265

Acknowledgments
These notes accompany the course Math 291 Intensive Linear Algebra and Multivariable Calculus
oered at Northwestern University. Anna Marie Bohmann, Paul Goerss, Ben Weinkove, Eric Zaslow
and Dylan Moreland suggested valuable mathematical and stylistic improvements. I am grateful to
the following students who diligently read the manuscript, oered suggestions, and reported errors:
Joseph Breen, Matthew Du, Alejandro Jensen, David Kim, Pallavi Pandey, Adam Ratner, Clayton
Shuttleworth, Alberto Takase, Melody Xu,Yang Yu and Kevin Zhao. The responsibility for any
remaining errors is entirely my own.
John M. Alongi
Chicago, IL
September 2013

iii

iv

CONTENTS

Chapter 1

Vectors
Vectors are fundamental to the study of linear algebra and multivariable calculus. In this chapter
we explore vectors from both geometric and coordinate viewpoints. The geometric perspective
provides intuition and motivation for the formal development in coordinates.

1.1

Geometric Vectors

Objective
Provide an intuitive introduction to geometric vectors and vector operations.
Some physical quantities such as displacement, velocity, acceleration, force and momentum
possess a magnitude and a direction. We refer to such quantities as vector quantities. It is natural
to represent a vector quantity by a directed line segment (or arrow) whose length is proportional
to the magnitude of the vector quantity, and whose direction indicates the direction of the vector
quantity.
A geometric vector is a directed line segment. See Figure 1.1. We denote geometric vectors
by boldface letters such as a or by an arrow over a letter1 such as ~a. As a directed line segment,

Figure 1.1: A Geometric Vector

!
a geometric vector has an initial point A and a terminal point B. The notation AB specifies a
1 Use

arrows over letters when notating vectors by hand to distinguish vectors from other mathematical objects.

CHAPTER 1. VECTORS

geometric vector by its initial point A and terminal point B. The term vector is Latin for carrier.
!
Imagine that the vector AB carries its initial point A to its terminal point B.
terminal point
B

!
AB

initial point

Figure 1.2: The Initial Point and Terminal Point of a Geometric Vector
Geometric vectors a and b are equal, denoted
a=b
if a and b have the same magnitude and direction. Geometric vectors do not have to consist of the
same set of points to be equal. For example, in Figure 1.3 the geometric vectors a and c are equal
because they have the same magnitude and direction, even though they are not the same directed
line segment. We regard a and c as the same vector. Consequently, we can move a geometric vector
in a plane or in space, and it is still the same vector as long as we do not alter its magnitude or
direction,. On the other hand, the vectors a and b in Figure 1.3 are not equal because they do not
have the same direction. Similarly, the vectors a and d are not equal.
b

a
c

Figure 1.3: Examples of Equal Geometric Vectors and Unequal Geometric Vectors
To add two geometric vectors a and b, draw a and b so that initial point of b coincides with
the initial point of a. The geometric vectors a and b determine a parallelogram. The vector sum
a + b is the diagonal of this parallelogram directed so that its initial point is at the common initial
point of a and b. This is the parallelogram rule for adding geometric vectors. See Figure 1.4.
From the parallelogram rule it is clear that
a+b=b+a
because the parallelogram and diagonal obtained by first drawing a and then drawing b are the
same, respectively, as the parallelogram and diagonal obtained by first drawing b and then drawing

1.1. GEOMETRIC VECTORS

a+b

b
a

Figure 1.4: The Vector Sum: Parallelogram Rule

a. Because the sum of geometric vectors is independent of the order of the vectors, we say that the
vector sum is commutative.
Alternatively, to add the geometric vectors a and b, draw a and b so that the initial point of b
coincides with the terminal point of a. The vector sum a + b is the vector whose initial point is the
initial point of a and whose terminal point is the terminal point of b. The vectors a, b and a + b
form the sides of a triangle. This is the triangle rule for adding geometric vectors. See Figure 1.5.
Comparing Figures 1.4 and 1.5 we see that the triangle rule and the parallelogram rule produce the

a+b

Figure 1.5: The Vector Sum: Triangle Rule

same vector sum a + b.
Using the triangle rule, notice that
(a + b) + c = a + (b + c)
See Figure 1.6. Because the sum of geometric vectors is independent of the grouping of the vectors,
we say that the vector sum is associative.
For each geometric vector a we denote the geometric vector with same magnitude as a but in
the opposite direction from a by a. The vector dierence of a and b is the geometric vector
a

def

b = a + ( b)

To subtract b from a, draw b and add a to b using the parallelogram rule. See Figure 1.7. The
initial point of a b is the terminal point of b, and the terminal point of a b is the terminal
point of a.

CHAPTER 1. VECTORS

(a + b) + c = a + (b + c)

b+c

a+b

Figure 1.6: Associativity of the Vector Sum

b
a
b

Figure 1.7: The Vector Dierence

For the purposes of this section a scalar is a real number. (In Section 1.3 we formally define the
term scalar in a way that allows complex numbers to be scalars under some circumstances.) We
denote scalars by letters in regular or italic font such as k.
To multiply a geometric vector a by a scalar k, draw a. If |k| 1, then expand the magnitude
of a by a factor of |k|. If |k| < 1, then contract the magnitude of a by a factor of k. If k < 0,
then reverse the direction of a. The scalar-vector product of k with a is the geometric vector ka
obtained from this procedure. See Figure 1.8.
Scalars arise in at least two ways. First, a scalar may describe a physical quantity such as
distance, speed, mass, charge and temperature that possesses a magnitude but not a direction. In
particular, the magnitude of the geometric vector is a scalar.
Second, a scalar may express the expansion, contraction, reflection or a combination of an
expansion/contraction and a reflection of a geometric vector. For example, represent a nonzero
velocity by a geometric vector v.
Multiplying v by the scalar 2 produces a geometric vector 2v that represents a velocity with
the same direction as v but with twice the magnitude as v. The geometric vector 2v is an
expansion of v.

1.1. GEOMETRIC VECTORS

ka (k > 1)

Figure 1.8: The Scalar-Vector Product

Multiplying v by the scalar 1/2 yields a geometric vector (1/2)v that represents a velocity
with the same direction as v but with half the magnitude as v. The geometric vector (1/2)v
is a contraction of v.
Multiplying v by the scalar

1 yields the geometric vector

def

1v =

that represents a velocity with the same magnitude as v but in the opposite direction from
v. The geometric vector 1v is a reflection of v.
Multiplying v by the scalar 1/2 produces a geometric vector ( 1/2)v that represents a
velocity with half the magnitude of v and in the opposite direction from v.
Concept
Vectors may represent physical quantities possessing magnitude and direction.
Skills
Add geometric vectors using the parallelogram rule.
Add geometric vectors using the triangle rule.
Subtract geometric vectors.

Multiply a geometric vector by a scalar.

Exercise
1. Copy the vectors in the figure, and draw the following vectors.
(a) a + b
(b)

(c) a

(d) 2a
(e)

1
2b

CHAPTER 1. VECTORS

Figure 1.9: Vectors for Exercise 1

1.2

Rectangular Coordinate Systems

Objective
Develop rectangular coordinate systems on lines and planes and in space.
A coordinate system uses numbers to uniquely determine the position of a point in a geometric object such as a line, a plane, or space. Coordinates allow us to describe geometric problems
algebraically and vice versa. In this section we provide an intuitive introduction to rectangular coordinate systems. A formal treatment of coordinates appears in Section 5.9. Later we study nonlinear
coordinate systems such as polar coordinates, cylindrical coordinates and spherical coordinates.

1.2.1

Rectangular Coordinates on a Line

We begin by establishing a one-to-one correspondence between the set of points on any line and
the set R of real numbers. A one-to-one correspondence between a line l and R is a relation that
associates exactly one real number to each point of l and associates exactly one point of l to each
real number.
Consider a line l. Choose a point O on l called the origin of the coordinate system. The point
O separates the line l into two rays. Choose one of the rays to be the positive ray of l and the other
to be the negative ray of l. There are two ways of choosing which ray is positive, each called an
orientation of l. Choose a point I on the positive ray to represent a unit length on the line. See
Figure 1.10.

Figure 1.10: A Rectangular Coordinate System on a Line

To assign a unique real number to each point, let P be a point on l. There exists a real number
a such that
!
!
OP = aOI
Define the coordinate of P to be a if P lies on the positive ray of l and
ray of l.

a if P lies on the negative

1.2. RECTANGULAR COORDINATE SYSTEMS

To assign a unique point to each real number, let a 2 R. If a

0, then define P to be the
!
!
point on the positive ray of l such that OP = aOI. If a < 0, then define P to be the point on the
!
!
negative ray of l such that OP = |a|OI. See Figure 1.11.
a
O

Figure 1.11: A Coordinate Correspondence on a Line

1.2.2

Rectangular Coordinates on a Plane

Next, we establish a one-to-one correspondence between the set of points in any plane and the set
R2 of ordered pairs of real numbers. Symbolically,
def

R2 = {(a1 , a2 ) : a1 , a2 2 R}
Consider a plane P. Choose a point O in P called the origin of the coordinate system. Choose
two perpendicular lines in P intersecting at O called the coordinate axes of the coordinate system. Choose an orientation and a unit of length for each axis so that each axis is in one-to-one
correspondence with the set of real numbers as in Subsection 1.2.1.
There are two possible orders for the coordinate axes. Choose an order for the coordinate axes,
calling the first axis the x-axis and calling the second axis the y-axis. Each order of the coordinate
axes is an orientation of the plane. If rotating the positive ray of the x-axis by 90 counterclockwise
transforms the positive ray of the x-axis into the positive ray of the y-axis, then the orientation is
positive. If rotating the positive ray of the x-axis by 90 clockwise transforms the positive ray of
the x-axis into the positive ray of the y-axis, then the orientation is negative.
The coordinate axes separate the plane into four regions called quadrants. We usually label
quadrants with Roman numerals I, II, III and IV, starting in the quadrant bounded by the positive
rays of the x- and y- axes and proceeding counterclockwise. See Figure 1.12.
To assign a unique ordered pair of real numbers to each point P in P, construct a line through
P perpendicular to the x-axis. Let a1 be the coordinate of the intersection of this line with the
x-axis. Construct a line through P perpendicular to the y-axis. Let a2 be the coordinate of the
intersection of this line with the y-axis. The ordered pair (a1 , a2 ) corresponds to the point P .
To assign a unique point to each ordered pair (a1 , a2 ) 2 R2 , let Q be the point with coordinate
a1 on the x-axis. Let R be the point with coordinate a2 on the y-axis. Construct a line m through
Q perpendicular to the x-axis, and construct a line n through R perpendicular to the y-axis. The
lines m and n intersect at a unique point P . The point P corresponds to the ordered pair (a1 , a2 ).
See Figure 1.13.
A rectangular coordinate system in the plane is often called a Cartesian coordinate system in
honor of the French philosopher, mathematician and writer Rene Descartes.2
2 Ren
e Descartes (15961650), the Father of Modern Philosophy, is perhaps best known for his statement, Cogito
ergo sum (English: I think, therefore I am).

CHAPTER 1. VECTORS
y

II

IV

III
P

Figure 1.12: A Rectangular Coordinate System on a Plane

y
P (a1 , a2 )
R a2

Q a1

Figure 1.13: A Coordinate Correspondence on a Plane

1.2.3

Rectangular Coordinates in Space

Finally, we define a one-to-one correspondence between the set of points in space and the R3 of
ordered triples of real numbers. Symbolically,
def

R3 = {(a1 , a2 , a3 ) : a1 , a2 , a3 2 R}
Choose a point O in space called the origin of the coordinate system. Choose three mutually
perpendicular lines in space intersecting at O called the coordinate axes of the coordinate system.
Choose a unit of length and an orientation for each axis. Each pair of coordinate axes determines a
coordinate plane of the coordinate system. Each coordinate plane is in one-to-one correspondence
with R2 as in Subsection 1.2.2.
There are eight possible orders for the coordinate axes. Choose an order for the coordinate
axes, calling the first axis the x-axis, calling the second axis the y-axis, and calling the third axis

1.2. RECTANGULAR COORDINATE SYSTEMS

the z-axis. Each ordered choice of coordinate axes corresponds to one of two possible orientations
of space. To determine the orientation of an ordered choice of coordinate axes in space we use the
right-hand rule:
Arrange the thumb, index, and middle fingers of your right hand so that they are perpendicular
to each other (with the index finger pointed straight) and so that the thumb points in the
direction of the positive ray of the x-axis and the index finger points in the direction of the
positive ray of y-axis.
If your middle finger points in the direction of the positive ray of the z-axis, then the order
of the coordinate axes is a positive orientation of space.
If your middle finger points in the direction of the negative ray of the z-axis, then the order
of the coordinate axes is a negative orientation of space.

Figure 1.14: The Right-Hand Rule

When we render a positively-oriented coordinate system in space on paper or on a blackboard we
usually draw the positive ray of the x-axis as if it were coming out of the page toward the reader
(or out of the blackboard toward the audience). See Figure 1.15.
We can label the coordinate planes according to the coordinate axes they contain. The coordinate plane containing the x- and y-axes is the xy-plane. The coordinate plane containing the xand z-axes is the xz-plane. The coordinate plane containing the y- and z-axes is the yz-plane.
The three coordinate planes separate space into eight orthants. The first orthant is the region
of space bounded by the half-planes of the coordinate planes containing the positive ray of each
coordinate axis. There does not appear to be a universal convention about how to number the
other seven orthants. See Figure 1.15.
To assign a unique ordered triple of real numbers to each point P in space, construct a plane
through P perpendicular to the x-axis. Let a1 be the coordinate of the intersection of this plane
with the x-axis. Construct a plane through P perpendicular to the y-axis. Let a2 be the coordinate
of the intersection of this plane with the y-axis. Construct a plane through P perpendicular to the
z-axis. Let a3 be the coordinate of the intersection of this plane with the z-axis. The ordered triple
(a1 , a2 , a3 ) corresponds to the point P .

10

CHAPTER 1. VECTORS
z

first octant
O
y
x

Figure 1.15: A Rectangular Coordinate System in Space

To assign a unique point to each ordered triple (a1 , a2 , a3 ) 2 R3 , let Q be the point with
coordinate a1 on the x-axis. Let R be the point with coordinate a2 on the y-axis. Let S be a
point with coordinate a3 on the z-axis. Construct a plane P through Q perpendicular to the xaxis, construct a plane Q through R perpendicular to the y-axis, and construct a plane R through
S perpendicular to the z-axis . The planes P, Q and R intersect at a point P . The point P
corresponds to the ordered triple (a1 , a2 , a3 ). See Figure 1.16.
Terms
origin
positive ray
negative ray
orientation
coordinate axes
coordinate planes
Concepts
The set of real numbers is in one-to-one correspondence with the points on a line.
The set of ordered pairs of real numbers is in one-to-one correspondence with the points in a
plane.
The set of ordered triples of real numbers is in one-to-one correspondence with the points in
space.

11

1.2. RECTANGULAR COORDINATE SYSTEMS

S
P (a1 , a2 , a3 )
O

a3

a1

a2

P
x
Q

Figure 1.16: A Coordinate Correspondence in Space

Skills
Find the coordinate of a point on a line equipped with a rectangular coordinate system.
Find the point on a line equipped with a rectangular coordinate system corresponding to a
given coordinate.
Find the coordinates of a point in a plane equipped with a rectangular coordinate system.
Find the point in a plane equipped with a rectangular coordinate system corresponding to a
given ordered pair of coordinates.
Find the coordinates of a point in space equipped with a rectangular coordinate system.
Find the point in space equipped with a rectangular coordinate system corresponding to a
given ordered triple of coordinates.
Exercises
1. Match each of the following rectangular coordinates with its corresponding point on the line
l. The unlabeled point corresponds to 1.

12

CHAPTER 1. VECTORS
(a) 2
(b)

1/2

2. Match each pair of rectangular coordinates with its corresponding point in Figure 2. The
unlabeled points correspond to 1 on each coordinate axis.
(a) (2, 3)
(b) ( 1, 2/3)
p
(c) (
2, 3/5)
(d) (e, /2)
y

x
S

Figure 1.17: The Rectangular Coordinate System for Exercise 2

3. Match each triple of rectangular coordinates with its corresponding point in Figure 3. The
unlabeled points correspond to 1 on each coordinate axis.
z

R
P

O
S

Figure 1.18: The Rectangular Coordinate System for Exercise 3

1.3. COORDINATE VECTORS

13

(a) (1, 2, 3)
(b) (0, 0, 1/2)
(c) ( 2, /2, 0)
(d) (2, 3/5, e)
4. Draw a rectangular coordinate system on a line, and plot the points with the following rectangular coordinates.3
(a)

2/3

(b)
5. Draw a rectangular coordinate system on a plane, and plot the points with the following
rectangular coordinates.
(a) (1, 1)
(b) (2, 0)
(c) (0, 3)

p
(d) ( 1/2, 3)
6. Draw a rectangular coordinate system in space, and plot the points with the following rectangular coordinates.
(a) (1, 1, 1)
(b) (2, 3, 0)
(c) (0, 0, 1)
(d) (0, 2, 3)
7. Identify each of the rectangular coordinate systems in Figure 7 as positively oriented or
negatively oriented.
8. Identify each of the rectangular coordinate systems in Figure 8 as positively oriented or
negatively oriented.

1.3

Coordinate Vectors

Objective
Define coordinate vectors, scalars, components of a vector, and equality of vectors.
In this we section we begin the formal development of vectors and scalars using the intuition
gained in the previous two sections about geometric vectors and rectangular coordinate systems.
3 Use

a straightedge when drawing figures.

14

CHAPTER 1. VECTORS
y

(a)

(b)

(c)

Figure 1.19: The Rectangular Coordinate Systems for Exercise 7

For each positive integer n denote the set of n-tuples of real numbers by Rn , and denote the set
of n-tuples of complex numbers by Cn . Symbolically,
def

Rn = {(a1 , . . . , an ) : a1 , . . . , an 2 R}
and

def

Cn = {(a1 , . . . , an ) : a1 , . . . , an 2 C}

Throughout this text it is convenient to define terms and state theorems in a way that makes
sense for both real numbers and complex numbers simultaneously. We use the symbol F to represent
either R or C when a term or theorem applies when replacing F by R and when replacing F by C.
Definition 1.3.1. A scalar is an element of F.
Definition 1.3.2. Let n be a positive integer. Define
def

Fn = {(a1 , . . . , an ) : a1 , . . . , an 2 F}
A coordinate vector (or vector) is an element of Fn .
We denote vectors by boldface letters. For example, we might define a coordinate vector a in
Fn by
def
a = (a1 , . . . , an )

15

1.3. COORDINATE VECTORS

z

x
O

O
y

x
z
(a)

(b)

(c)
z

x
O

O
y

x
z

z
(d)

(e)

(f)

Figure 1.20: The Rectangular Coordinate Systems for Exercise 8

Often it is useful to regard a as a column such as
2
3
a1
6 .. 7
4 . 5
an
Example 1.3.3.

p
(a) (0, 1, 2/3, 5) is a vector in R4 .

1
(b)
is a vector in C2 .
i
We use n-tuple notation when interpreting vectors geomtetically as points, and we use column
notation when interpreting vectors geometrically as directed line segments. Because
it is useful to

distinguish between rows and columns, the notation a will not denote a row such as a1 an .
def

Definition 1.3.4. The components of a vector a = (a1 , . . . , an ) are the scalars a1 , . . . , an .

Example 1.3.5. The components of the vector (1, 2, 3) in R3 are the real numbers 1,

2 and 3.

Two vectors are equal exactly when their corresponding components are equal.
def

def

Definition 1.3.6. Vectors a = (a1 , . . . , an ) and b = (b1 , . . . , bn ) in Fn are equal, denoted a = b,

if aj = bj for all j = 1, . . . , n.

16

CHAPTER 1. VECTORS
y

(a1 , a2 )

a2

a1

Figure 1.21: A Vector in R2

z
a3

(a1 , a2 , a3 )

a2

a1
x

Figure 1.22: A Vector in R3

Example 1.3.7. The vectors (1, 2, 3) and (2, 1, 3) in R3 are not equal even though their components
are equal because not all of their corresponding components are equal.
Equality of vectors enjoys the same properties as equality of scalars.
Theorem 1.3.8.
(i) (Reflexivity) For all a 2 Fn , a = a.
(ii) (Symmetry) For all a, b 2 Fn , if a = b, then b = a.
(iii) (Transitivity) For all a, b, c 2 Fn , if a = b and b = c, then a = c.

Proof.

(i) Exercise.
(ii) Let a, b 2 Fn . Let a1 , . . . , an be the components of a, and let b1 , . . . , bn be the components
of b. Assume that a = b. For each j = 1, . . . , n, aj = bj . Because equality of scalars is
symmetric, bj = aj for each j = 1, . . . , n. Therefore, b = a.
(iii) Exercise.

1.3. COORDINATE VECTORS

17

Terms
scalar
coordinate vector
components
equal vectors
Concepts
Scalars are real or complex numbers.
A coordinate vector is an n-tuple of scalars called the components of the coordinate vector.
Two vectors are equal exactly when their corresponding components are equal.
Skills
Use vector notation correctly.
Identify the components of a coordinate vector.
Interpret coordinate vectors geometrically as points and as directed line segments.
Determine whether two coordinate vectors are equal.
Exercises
1. Draw a rectangular coordinate system on a plane, and draw the following vectors.
(a) (1, 1)
(b) (0, 3)

2
(c)
0

1/2
p
(d)
3
2. Draw a rectangular coordinate system in space, and draw the following vectors.
(a) ( 1, 1, 1)
(b) (2, 3, 0)
2 3
0
(c) 4 0 5
1

18

CHAPTER 1. VECTORS
2

(d) 4

3
0
2 5
0

3. For what values of x and y is (0, x) = (ln y, 2)?

2
3 2
3
arctan y
1
5 = 4 ex 5?
1
4. For what values of x, y and z is 4
cos(/6)
z
5. Prove Theorem 1.3.8(i) and (iii).

1.4

The Vector Sum

Objective
Develop the vector sum and its properties algebraically.
To add two vectors we add corresponding components.
def

def

Definition 1.4.1. The vector sum of a = (a1 , . . . , an ) 2 Fn and b = (b1 , . . . , bn ) 2 Fn is the

vector
def
a + b = (a1 + b1 , . . . , an + bn )
The operation of forming the a vector sum is called vector addition.
Example 1.4.2.

Express the sum of an

n-tuple and a column
vector as an n-tuple.

(a) (1
2

2i, 3i) + ( 4 + 5i, 7) = ( 3 + 3i, 7 + 3i)

3 2
3 2
3
e
e
0
5=4
5
2/3
(b) 4 1 5 + 4 1/3
p
p

2
+ 2

3
(c) (2, 1) +
= ( 1, 4)
5

In Section 1.1 we introduced the sum of two geometric vectors intuitively using the parallelogram
rule and the triangle rule. Figures 1.23 and 1.24 show that the formal definition of the vector sum
in Definition 1.4.1 agrees with the parallelogram rule and the triangle rule., respectively.
The next theorem expresses the most important properties of the vector sum.
Theorem 1.4.3.
(i) (Commutativity) a + b = b + a for all a, b 2 Fn .
(ii) (Associativity) a + (b + c) = (a + b) + c for all a, b, c 2 Fn .
(iii) (Existence of a Unique Additive Identity) There is a unique vector 0 2 Fn such that a + 0 = a
for all a 2 Fn .

19

1.4. THE VECTOR SUM

y

a2 + b2
b2

a+b
b
a2

a
a1 + b1

a1

b1

Figure 1.23: The Vector Sum: Parallelogram Rule

a2 + b2

a+b

b2

a2

b1

a1

a1 + b1

Figure 1.24: The Vector Sum: Triangle Rule

(iv) (Existence of Unique Additive Inverses) For each a 2 Fn there is a unique vector
such that a + ( a) = 0.
Proof.
def

def

(i) Let a = (a1 , . . . , an ) and b = (b1 , . . . , bn ) be vectors in Fn .

a + b = (a1 , . . . , an ) + (b1 , . . . , bn )

(ii) Exercise.

Definitions of a and b

= (a1 + b1 , . . . , an + bn )

Definition of the Vector Sum

= (b1 + a1 , . . . , bn + an )

Commutativity of Addition of Scalars

=b+a

Definition of the Vector Sum

a 2 Fn

20

CHAPTER 1. VECTORS
def

def

(iii) Let a = (a1 , . . . , an ) be a vector in Fn . Define 0 = (0, . . . , 0).

a + 0 = (a1 , . . . , an ) + (0, . . . , 0)

Definitions of a and 0

= (a1 + 0, . . . , an + 0)

Definition of the Vector Sum

= (a1 , . . . , an )

Additive Identity Property of Scalars

=a

Definition of a

To prove that 0 is the unique vector with this property, assume that z 2 Fn and a + z = a
for all a 2 Fn . In particular,
0+z=0
On the other hand, part (i) implies that
0+z=z+0=z
Thus, z = 0. Therefore, 0 is the unique vector in Fn such that a + 0 = a for all a 2 Fn .
(iv) Exercise.

By the associativity of the vector sum (Theorem 1.4.3 (ii)) the meaning of the notation a + b + c
is unambiguous because the sum is the same no matter where we insert parentheses.
Theorem 1.4.4. (Generalized Associativity) For each integer p
a1 + (a2 + + ap ) = (a1 + + ap

3,
1)

+ ap

for all a1 , . . . , ap 2 Fn .
Proof. Exercise. Use the Principle of Mathematical Induction.
For all a1 , . . . , ap 2 Fn , the meaning of a1 + + ap is unambiguous by Theorem 1.4.4. This
justifies the use of the summation notation
p
X
j=1

def

aj = a1 + + ap

Definition 1.4.5. The zero vector in Fn is the vector

def

0 = (0, . . . , 0)
Definition 1.4.6. The vector dierence of a vector a 2 Fn and a vector b 2 Fn is the vector
a
Express the difference
of two n-tuples as a
column vector and
express the difference of
an n-tuple and column
vector as an n-tuple.

Example 1.4.7.

def

b = a + ( b)

21

1.4. THE VECTOR SUM

(a) (1, 2i, 3 + 4i)

(b) ( 1, 2)
(c)

0
1

3
5
2
3

3
1 + 4i
2i 5
( 4i, 0, 3) = 4
4i
= ( 4, 7)
=

2
2

Using the vector dierence and a rectangular coordinate system we can represent a geometric
vector by a coordinate vector. See Figure 1.25.
!
P1 P2

P2

P1

Figure 1.25: The Vector Between Two Points

!
Example 1.4.8. Represent the geometric vector P1 P2 with initial point P1 (2, 3) and terminal
point P2 ( 5, 7) by a coordinate vector.
def

def

Solution. Let p1 = (2, 3) and p2 = ( 5, 7) be the position vectors of the points P1 and P2 ,
!
respectively. Represent the geometric vector P1 P2 by the vector dierence

7
p2 p1 = ( 5, 7) (2, 3) =
10
N
Terms
vector sum
zero vector
vector dierence
Concepts
The vector sum is a commutative and associative operation.
The vector sum has a unique identity called the zero vector.
Every vector has a unique additive inverse.

22

CHAPTER 1. VECTORS

Skills
Find the sum of two vectors.
Find the dierence of two vectors.
Draw a vector sum using the triangle rule and the parallelogram rule.
Draw a vector dierence.
Represent a geometric vector by a coordinate vector.
Exercises
1. Find each vector sum or dierence.
(a) (3, 1) + ( 1, 7)
2
3
3/5
(b) (0, 1, 2) + 4 p 5
7
2 3 2
3
1
5
6 2 7 6 1 7
7 6
7
(c) 6
4 3 5+4 2 5
4
0
(d) (1

2i, 3i) + ( 5, 7 + 11i)

(e) (1, 2)
(f) (1, 0, 2)
(g)

2/3

(5/3, e)
2 3
2
4 1 5
7

def

p
5 + i 13
0

def

2. Let a = (1, 2) and b = ( 2, 5).

(a) Draw a + b using the triangle rule.
(b) Draw a + b using the parallelogram rule.
(c) Draw a

b.
def

3. Let n be a positive integer. Let a = (1, 3, 5, . . . , 2n

def

1) and b = (2, 4, 6, . . . , ( 1)n+1 2n).

(a) Find a + b.
(b) Find a

b.

!
4. Represent the geometric vector P1 P2 with initial point P1 and terminal point P2 by a coordinate vector.

23

1.5. THE SCALAR-VECTOR PRODUCT

(a) P1 (2, 4), P2 (3, 9)
(b) P1 (4, 1, 0), P2 (3, 5, 3)
5. Justify each step in the following proof of Theorem 1.4.3 (ii).
def

def

def

Proof. Let a = (a1 , . . . , an ), b = (b1 , . . . , bn ), and c = (c1 , . . . , cn ) be vectors in Fn .

a + (b + c) = a + ((b1 + c1 , . . . , bn + cn ))
= (a1 + (b1 + c1 ), . . . , an + (bn + cn ))
= ((a1 + b1 ) + c1 , . . . , (an + bn ) + cn ))

Definition 1.4.1
(a)
(b)

= (a1 + b1 , . . . , an + bn ) + c

(c)

= (a + b) + c

(d)

Therefore, for all a, b, c 2 Fn , a + (b + c) = (a + b) + c.

6. Prove Theorem 1.4.3 (iv).
7. Prove Theorem 1.4.4.
8. Prove that for all a, b, c 2 Fn , if a = b, then a + c = b + c and c + a = c + b.
9. Justify each step in the following proof that for all a, b 2 Fn , if a + b = a, then b = 0.
Proof. Let a, b 2 Fn . Assume that a + b = a.
a+b=a

Hypothesis

a + (a + b) =

a+a

Exercise 8

( a + a) + b =

a+a

(a)

(a + ( a)) + b = a + ( a)

Theorem 1.4.3 (i)

0+b=0

(b)

b+0=0

(c)

b=0

(d)

Therefore, for all a, b 2 Fn , if a + b = a, then b = 0.

1.5

The Scalar-Vector Product

Objective
Develop the scalar-vector product and its properties algebraically.
Define parallel vectors and the concept of direction using the scalar-vector product.

24

CHAPTER 1. VECTORS

Definition 1.5.1. The scalar-vector product of a scalar k and a vector a = (a1 , . . . , an ) is the
vector
def
ka = (ka1 , . . . , kan )
The operation of forming the product of a scalar with a vector is called scalar multiplication.
Example 1.5.2.
(a) 3(1, 2) = (3, 6)
2
3 2
1
6 3/5 7 6
7=6
(b) 2 6
4
5 4
4 7i

3
2
6/5 7
7
5
2
8 + 14i

ka2
ka
a2

ka1

a1

Figure 1.26: The Scalar-Vector Product

Figure 1.26 shows that the formal definition of the scalar-vector product in Definition 1.5.1
agrees with the intuitive approach to scalar multiplication in Section 1.1.
The next theorem establishes the most important properties of the scalar-vector product.
Theorem 1.5.3.
(i) (Associativity) For all a 2 Fn and all scalars k and l,
k(la) = (kl)a = l(ka)
(ii) (Existence of a Multiplicative Identity) For all a 2 Fn ,
1a = a
(iii) (Distributivity of the Scalar-Vector Product over the Scalar Sum) For all a 2 Fn and all
scalars k and l,
(k + l)a = ka + la

25

1.5. THE SCALAR-VECTOR PRODUCT

(iv) (Distributivity of the Scalar-Vector Product over the Vector Sum) For all a, b 2 Fn and all
scalars k,
k(a + b) = ka + kb
Proof.
(i) Exercise.
(ii) Exercise.
def

(iii) Let a = (a1 , . . . , an ) be a vector in Fn . Let k and l be scalars.

(k + l)a = ((k + l)a1 , . . . , (k + l)an )

Definition of the Scalar-Vector Product

= (ka1 + la1 , . . . , kan + lan )

Distributivity of Multiplication over Addition of Scalars

= (ka1 , . . . , kan ) + (la1 , . . . , lan )

Definition of the Vector Sum

= ka + la

Definition of the Scalar-Vector Product

def

def

(iv) Let a = (a1 , . . . , an ) and b = (b1 , . . . , bn ) be vectors in Fn . Let k be a scalar.

k(a + b) = k((a1 , . . . , an ) + (b1 , . . . , bn ))

Definitions of a and b

= k(a1 + b1 , . . . , an + bn )

Definition of the Vector Sum

= (k(a1 + b1 ), . . . , k(an + bn ))

Definition of the Scalar-Vector Product

= (ka1 + kb1 , . . . , kan + kbn )

Distributivity of Multiplication over Addition of Scalars

= (ka1 , . . . , kan ) + (kb1 , . . . , kbn )

Definition of the Vector Sum

= ka + kb

Definition of the Scalar-Vector Product

Theorem 1.5.4.
(i) For all a 2 Fn , 0a = 0.

(ii) For all scalars k, k0 = 0.

(iii) For all a 2 Fn , ( 1)a =

a.

Proof.

(i) Exercise.
(ii) Exercise.
def

(iii) Let a = (a1 , . . . , an ) be a vector in Fn .

a + ( 1)a = ((a1 , . . . , an ) + ( 1a1 , . . . , 1an ))

Definition of the Scalar-Vector Product

= ((a1 , . . . , an ) + ( a1 , . . . , an ))

Additive Inverse Property of Scalars

= (a1 + ( a1 ), . . . , an + ( an ))

Definition of the Vector Sum

= (0, . . . , 0)

Existence of Additive Inverses

=0

Definition of the Zero Vector

26

CHAPTER 1. VECTORS
By Theorem 1.4.3 (iv) there is a unique vector
( 1)a = a.

1.5.1

a 2 Fn so that a + ( a) = 0. Therefore,

Parallel Vectors

Two vectors in Fn are parallel if one is a scalar multiple of the other.

Definition 1.5.5. Vectors a, b 2 Fn are parallel, denoted a k b, if there is a scalar k such that
b = ka

or

a = kb

Because 0 = 0a for all a 2 Fn , the zero vector in Fn is parallel to every vector in Fn .4

Definition 1.5.6. A vector a 2 Rn is in the direction of b 2 Rn if there is a nonnegative real
scalar k such that
b = ka
or
a = kb
The zero vector in Rn is in the direction of every vector in Rn . If a is in the direction of b, then
a is parallel to b. Notice that Definition 1.5.6 applies only to vectors with real components.

Figure 1.27: Parallel Vectors

In Figure 1.27 the vectors a, b and c are parallel to one another. However, a and b are in the
same direction while a and c are not in the same direction.
def

def

Example 1.5.7. The vectors a = (1, 2, 3) and b = ( 2, 4, 6) in R3 are parallel because

( 2, 4, 6) =

2(1, 2, 3)

However, a is not in the direction of b.

The next proposition provides an important characterization of parallel vectors that we will
generalize with our development of linear dependence in Chapter 3.
4 Some texts define vectors to be parallel only if one of the vectors is a nonzero scalar multiple of the other. Such
a definition does not allow the zero vector to be parallel to any vector other than itself.

1.5. THE SCALAR-VECTOR PRODUCT

27

Proposition 1.5.8. Vectors a, b 2 Fn are parallel if and only if there are scalars k1 and k2 at least
one of which is nonzero such that
k1 a + k2 b = 0
Proof. Exercise.
Terms
scalar-vector product
parallel vectors
vector in the direction of another
Concepts
The scalar-vector product is an associative operation.
The scalar-vector product has an identity element.
The scalar-vector product distributes over both scalar addition and vector addition.
Two vectors are parallel if one is a scalar multiple of the other.
Two vectors are in the same direction if one is a nonnegative real scalar multiple of the other.
Vectors a, b 2 Fn are parallel if and only if there are scalars k1 and k2 at least one of which
is nonzero such that
k1 a + k2 b = 0
Skills
Find the product of a scalar with a vector.
Represent the product of a scalar with a vector graphically.
Determine whether two vectors are parallel.
Determine whether a vector is in the direction of another.
Exercises
1. Find each scalar-vector product.
(a) 2( 3, 5)
(b)

2
3 (0,

p
1, 5)
3

0
6 1 7
7
(c) 6
4 2/3 5
p
5

28

CHAPTER 1. VECTORS
(d) i(1

2i, 3i)

2. Simplify.
(a) (8, 9) + 3( 1, 2)
2

3
5
(b) 12 (8, 4, 1) + 2 4 7 5
1/4
(c)

2 ((2, 0, 1)

6(1/2, 4, 1))

def

3. Let a = (1, 2).

(a) Draw 3a.
(b) Draw

3a.

(c) Draw 1a.

(d) Draw 0a.
4. Prove or disprove: If a 2 Fn , c 2 F and ca = a, then c = 1.
5. Prove Theorem 1.5.3 (i).
6. Justify each step in the following proof of Theorem 1.5.3 (ii).
def

Proof. Let a = (a1 , . . . , an ) 2 Fn .

1a = (1a1 , . . . , 1an )

(a)

= (a1 , . . . , an )

(b)

=a

Definition of a

Therefore, for all a 2 Fn , 1a = a.

7. Prove Theorem 1.5.4 (i).
8. Prove Theorem 1.5.4 (ii).
9. Determine whether each of the following statements is true or false. If the statement is true,
then prove it. If the statement is false, then provide a counterexample. Let a, b and c be
vectors in Fn .
(a) If a k b, then b k a.

(b) If a is in the direction of b, then b is in the direction of a.

(c) If a k b and b k c, then a k c.

(d) If a is in the direction of b, and b is in the direction of c, then a is in the direction of c.

10. Prove that (1, 2) and (5, 10) parallel vectors.

29

1.6. LINES

11. Let a, b 2 Fn . Prove that a k b if and only if either a = 0 or there exists a scalar k such that
b = ka.
12. Let a, b, c 2 Fn . Prove that if b is parallel to a and c is parallel to a, then b + c is parallel
to a.
13. Prove Proposition 1.5.8.

1.6

Lines

Objectives
Define lines and line segments in Fn .
Develop vector equations, sets of parametric equations and standard linear equations for lines.
tv
v
p + tv

l
p

Figure 1.28: A Line

Let v be a nonzero vector in R3 . The set of all scalar multiples tv is what we intuitively call a
line in R3 .
Definition 1.6.1. The line l in Fn containing p 2 Fn and parallel to a nonzero vector v 2 Fn is
def

l = {x 2 Fn : x

p = tv for some scalar t}

Definition 1.6.2. A vector equation for a line in Fn containing p 2 Fn and parallel to a

nonzero vector v 2 Fn is
x p = tv
for all scalars t.
Example
1.6.3. Find a vector equation for the line in R2 containing (1, 2) and parallel to

3/5
v=
.

Solution. A vector equation is

x
for all scalars t.

(1, 2) = t

3/5

30

1.6.1

CHAPTER 1. VECTORS

Parametric Equations for a Line

Let l be the line in Fn containing p 2 Fn and parallel to a nonzero vector v 2 Fn . A vector

x = (x1 , . . . , , xn ) lies on the line l if and only if there is a scalar t so that
p = tv

x
2

v1
6 .. 7
Define p = (p1 , . . . , pn ) and v = 4 . 5. Then the vector x lies on the line l if and only if there
vn
is a scalar t so that
2
3
v1
6
7
(x1 , . . . , xn ) (p1 , . . . , , pn ) = t 4 ... 5
def

def

2
6
4

x1

..
.

xn

p1

pn

vn

3
tv1
7
6 . 7
5 = 4 .. 5
tvn

Since vectors are equal exactly when their corresponding components are equal,
8
>
< x1 p1 = tv1
..
.
>
:
xn pn = tvn
def

Definition 1.6.4. Parametric 2

equations
for the line containing p = (p1 , . . . , pn ) 2 Fn and
3
v1
def 6
7
parallel to a nonzero vector v = 4 ... 5 2 Fn are
vn

8
>
< x1

for all scalars t.

>
:

xn

p1
pn

=
..
.

tv1

= tvn

Example
1.6.5. Find parametric equations for a line in R2 containing (1, 2) and parallel to

3/5
v=
.

Solution. Parametric equations are

x1 1 =
x2 + 2 =

3t/5
t
N

31

1.6. LINES

Example 1.6.6. Find parametric equations for a line in R3 containing (0, 1, 2) and ( 3, 5, 7).
Solution. A nonzero vector parallel to the line is
(0, 1, 2)

( 3, 5, 7) = 4

Therefore, parametric equations for the line are

8
=
< x
y+1 =
:
z 2 =

3
3
6 5
9

3t
6t
9t
N

for all scalars t.

Alternative Solution. Another nonzero vector parallel to the line is

2
3
3
( 3, 5, 7) (0, 1, 2) = 4 6 5
9
Therefore, parametric equations for the line are
8
< x+3 =
y 6 =
:
z+9 =

3t
6t
9t
N

for all scalars t.

Parametric equations for a line are never unique because you can choose any point on the line
and any vector parallel to the line to produce parametric equations for the line.

1.6.2

Linear Equations in Two Variables

Definition 1.6.7. A linear equation in two variables is an equation of the form

ax + by = c
where a, b and c are scalars so that either a or b is nonzero.
Example 1.6.8. Show that the set of (x, y) 2 R2 satisfying the linear equation
x

2y = 3

is a line.
Solution. Solving x

2y = 3 for x we obtain
x = 2y

def

Define t = y. Then x = 2t

3. Thus, parametric equations for the line are

x 3 = 2t
y
= t

32

CHAPTER 1. VECTORS

Alternative Solution. Solve x

2y = 3 for y to obtain
1
3
x+
2
2

y=
def

Define t = x. Then y =

1
2x

+ 32 . Thus, parametric equations for the line are

(

x
y

3
2

t
1
t
2

for all scalars t.

More generally, the set of vectors in a plane whose components satisfy a given standard linear
equation
ax + by = c
is a line.
Theorem 1.6.9. For all scalars a, b and c so that a and b are not both zero, the set of vectors
(x, y) in F2 whose components satisfy the linear equation
ax + by = c
is a line.
Proof. Let a, b and c be scalars so that a and b are not both zero.
Case I. a 6= 0
Solving ax + by = c for x we obtain
x=

c
a

b
y
a

def

Define t = y. Then the set of vectors (x, y) 2 F2 satisfying ax + by = c is exactly the set of vectors
that satisfy the set of parametric equations
(

x
y

c
a

=
=

b
t
a
t

for all scalars t.

Case II. a = 0
Because a and b are not both zero, b 6= 0. Solving ax + by = c for y we obtain
y=

c
b

33

1.6. LINES
def

Define t = x. Then the set of vectors (x, y) 2 F2 satisfying ax + by = c is exactly the set of vectors
that satisfy the set of parametric equations
(
x
= t
c
= 0
y
b
for all scalars t.
In each case the set of vectors (x, y) 2 F2 satisfying the linear equation ax + by = c satisfies a
set of parametric equations for a line.

1.6.3

Line Segments

p
q
Figure 1.29: A Line Segment

Definition 1.6.10. The line segment

def

= {x 2 Rn : x

in Rn with endpoints p, q 2 Rn is
p = t(q

p) for some t 2 [0, 1]}

Example 1.6.11. Find a vector equation for the line segment in R3 with endpoints (0, 1, 2) and
( 3, 5, 7). Specify the domain of the parameter in your parametric equations.
Solution. A nonzero vector parallel to the line segment is
( 3, 5, 7)

(0, 1, 2) = 4

Therefore, a vector equation for the line segment is

x
The domain of the parameter t is [0, 1].
Terms

(0, 1, 2) = t 4

3
3
6 5
9

3
3
6 5
9

34

CHAPTER 1. VECTORS
line
vector equation of a line
parametric equations of a line
linear equation in two variables
line segment

Concept
The line in Fn through p and parallel to a nonzero vector v is the set of vectors which dier
from p by a scalar multiple of v.
Skills
Find vector and parametric equations of the line containing p 2 Fn and parallel to a nonzero
vector v 2 Fn .
Find vector and parametric equations of a line containing p1 and p2 2 Fn .
Find vector and parametric equations of a line in F2 given by a linear equation in two variables.
Exercises
1.
Find vector and parametric equations for the line in R2 containing (2, 1) and parallel to
1
.
7
2. Find vector and parametric equations for the line in R3 containing (1, 4, 5) and (2, 4, 1).
3. 2
Find vector
and parametric equations for the line in R4 containing (1, 2, 0, 4) and parallel to
3
2
6 5 7
6
7
4 3 5.
7
p
4. Find vector and parametric equations for the line in R5 containing (9, , 1, 5, 2) and ( 1, 1, 2, 7, 1).
3
5. Find two dierent
2
3 sets of parametric equations for the line in R containing ( 1, 7, 3) and
2
parallel to 4 1 5.
5

6. Find two dierent sets of parametric equations for the line in R3 containing (5, 3, 4) and
(0, 1, 9).
7. Find vector and parametric equations for the lines in R2 with the following linear equations
in two variables.

35

1.7. PLANES
(a) 2x 3y = 5
(b) x = 1
(c) 2y = 3

8. (a) Prove that for all real numbers m and b, the set of vectors (x, y) in R2 whose components
satisfy the slope-intercept equation
y = mx + b
is a line.
(b) Find vector and parametric equations for the line in R2 with slope-intercept form y =
2x 3.
9. Find a vector equation for the line segment with endpoints (1, 3, 5) and ( 2, 4, 6). Specify
the domain of the parameter in your parametric equations.

1.7

Planes

Objective
Define planes and parallelograms in Fn algebraically, and obtain sets of parametric equations
and standard linear equations for planes.
Definition 1.7.1. The plane P in Fn containing p 2 Fn and parallel to the non-parallel vectors
a and b in Fn is
def

P = {x 2 Fn : x

p = sa + tb for some scalars s and t}

tb

sa + tb

sa
p + sa + tb

Figure 1.30: A Plane

Definition 1.7.2. A vector equation for a plane in Fn containing p 2 Fn and parallel to the
non-parallel vectors a and b in Fn is
x
for all scalars s and t.

p = sa + tb

36

CHAPTER 1. VECTORS

Example 1.7.3. Find a vector equation for the plane in R3 containing (1, 2, 3), ( 4, 5, 6) and
(7, 8, 9).
Solution. Two non-parallel vectors parallel to the plane are
2
( 4, 5, 6) = 4

(1, 2, 3)

and

(1, 2, 3)
Therefore, a vector equation for the plane is
x
for all scalars s and t.

1.7.1

3
5
7 5
9

(7, 8, 9) = 4
2

(1, 2, 3) = s 4

3
6
6 5
6

3
2
5
7 5 + t4
9

3
6
6 5
6

Parametric Equations for a Plane

Let P be the plane in Fn containing p 2 Fn and parallel to the non-parallel vectors a and b in Fn .
def
A vector x = (x1 , . . . , , xn ) lies in the plane P if and only if there are scalars s and t so that
p = sa + tb
2
3
a1
b1
def
def 6
def 6
7
7
Define p = (p1 , . . . , pn ), a = 4 ... 5, and b = 4 ... 5. Then the vector x lies in the plane P if
an
bn
and only if there are scalars s and t so that
2
3
2
3
a1
b1
6
7
6
7
(x1 , . . . , xn ) (p1 , . . . , , pn ) = s 4 ... 5 + t 4 ... 5
x

2
6
4

x1

xn

..
.

p1

pn

7
6
5 = 4

an

sa1 + tb1
7
..
5
.
san + tbn

bn

Since vectors are equal exactly when their corresponding components are equal,
8
>
< x1 p1 = sa1 + tb1
..
.
>
:
xn pn = san + tbn

37

1.7. PLANES
def

n
Definition 1.7.4. Parametric equations for
p = (p1 , . . . , pn ) 2 Fn
3 plane in F2 containing
3
2 the
a1
b1
def 6
def 6
7
7
.
and parallel to the non-parallel vectors a = 4 .. 5 and b = 4 ... 5 in Fn are
an
bn
8
>
< x1 p1 = sa1 + tb1
..
.
>
:
xn pn = san + tbn

for all scalars s and t.

Example 1.7.5. Find parametric equations for the plane in R3 containing (1, 2, 3), ( 4, 5, 6)
and (7, 8, 9).
Solution. Two non-parallel vectors parallel to the plane are
2
( 4, 5, 6) = 4

(1, 2, 3)

and
(1, 2, 3)

(7, 8, 9) = 4

Therefore, parametric equations for the plane

8
< x 1
y+2
:
z 3

are
=
=
=

3
5
7 5
9

3
6
6 5
6

5s 6t
7s + 6t
9s 6t

for all scalars s and t.

Parametric equations for a line are never unique because you can choose any point on the plane
and any pair of non-parallel vectors parallel to the plane to produce a set of parametric equations
for the plane.

1.7.2

Linear Equations in Three Variables

Definition 1.7.6. A linear equation in three variables is an equation of the form

ax + by + cz = d
where a, b, c and d are scalars so that a, b and c are not all zero.
Example 1.7.7. Show that the set of (x, y, z) 2 R3 satisfying the equation
x
is a plane.

2y + 3z = 5

38

CHAPTER 1. VECTORS

Solution. Solve x 2y + 3z = 5 for any one of the variables in terms of the other two. Solving for
x in terms of y and z,
x = 5 + 2y 3z
def

def

Define s = y and t = z. Then

x

5 = 2s

3t

So, a set of parametric equations is

8
< x
:

5
y
z

= 2s 3t
=
s
=
t

for all scalars s and t.

Thus, the set of (x, y, z) 2 R3 that satisfy the linear equation x 2y + 3z = 5 is exactly the set
of (x, y, z) that satisfy a set of parametric equations for a plane in R3 .
N
Theorem 1.7.8. For all scalars a, b, c and d such that a, b and c are not all zero, the set of vectors
(x, y, z) in F3 whose components satisfy the linear equation
ax + by + cz = d
is a plane.
Proof. Exercise.

1.7.3

Parallelograms

Definition 1.7.9. The parallelogram in Rn with vertex p 2 Rn and edges v1 , v2 2 Rn is

def

P = {x 2 Rn : x

p = sv1 + tv2 for some s, t 2 [0, 1]}

See Figure 1.31

3
Example
1.7.10.
2 3
2 3 Find a vector equation for the parallelogram in R with vertex (2, 7, 6) and edges
2
6
4 5 5 and 4 7 5. Specify the domain of each parameter in your vector equation.
5
7

Solution. The vertex of the parallelogram is (2, 7, 6). The edges of the parallelogram are
2 3
2 3
2
6
4 5 5 and 4 7 5
5
7
Therefore, a vector equation for this parallelogram is
2 3
2 3
2
6
x (2, 7, 6) = s 4 5 5 + t 4 7 5
5
7
The parameters are s and t. Each has domain [0, 1].

39

1.7. PLANES
z

v2
v1

Figure 1.31: A Parallelogram in R3

Terms
plane
vector equation of a plane
parametric equations of a plane
linear equation in three variables
parallelogram
Concept
The plane in Fn through p and parallel to non-parallel vectors v1 and v2 is the set of vectors
which dier from p by a sum of scalar multiples of v1 and v2 .
Skills
Find vector and parametric equations for the plane containing p 2 Fn and parallel to nonparallel vectors a and b in Fn .
Find vector and parametric equations for a plane containing p1 , p2 and p3 in Fn .

40

CHAPTER 1. VECTORS
Find vector and parametric equations for a plane in F3 given by a linear equation in three
variables.

Exercises
3
1. Find parametric
equations
2
3
2 for the
3 plane in R containing the point ( 1, 2, 7) and parallel to
2
1
the vectors 4 3 5 and 4 0 5.
1
5

2. Find vector and parametric equations for a plane in R3 containing (3, 1, 2), (2, 0, 5) and
(1, 2, 4).
3. Find two dierent sets of parametric equations for a plane in R3 containing (1, 2, 3), (0, 5, 7)
and ( 8, 6, 4).
4. Find vector and parametric equations for planes in R3 with the following linear equations in
three variables.
(a) 2x
(b) x
(c) 3y
(d) x

3y + 5z =

2y = 1
5z = 1
5z = 1

(e) x = 1
(f) 2y = 3
(g) 3z = 5
5. (a) Prove Theorem 1.7.8.
(b) Find vector and parametric equations for a plane in R3 with standard linear equation
2x

3y + 5z = 7

6. Find vector
parametric
equations
for the parallelogram in R3 with vertex ( 1, 2, 3)
2 and 3
2
3
4
7
and edges 4 5 5 and 4 8 5. Specify the domain for each parameter in your parametric
6
9
equations.

1.8

Linear Combinations

Objective
Introduce linear combinations of vectors in Fn and their elementary properties.

41

1.8. LINEAR COMBINATIONS

We have introduced two operations on vectors: addition and scalar multiplication. In the
previous two sections we defined lines and planes using vector addition and scalar multiplication.
Vectors formed using only the operations of vector addition and scalar multiplication are called
linear combinations. Linear algebra is the study of linear combinations.
Definition 1.8.1. A linear combination of vectors a1 , . . . , ap 2 Fn is a vector
p
X

kj aj

j=1

where k1 , . . . , kp are scalars called the coefficients5 of the linear combination.

8
Example 1.8.2. The vector
is a linear combination of the vectors
4
because

8
1
3
=3
+2
4
2
1

1
2

and

3
1

See Figure 1.32.

3
1

3
1
x

1
2

1
2

9
4

=2

3
1

+3

1
2

Figure 1.32: The Linear Combination of Example 1.8.2

Linear combinations allow elegant descriptions of lines and planes. Recall that a vector equation
for a line l in Fn containing p and parallel to a nonzero vector v is
x

p = tv

That is, a vector x lies on l exactly when x diers from p by a linear combination of v.
Similarly, recall that a vector equation for a plane P in Fn containing p and parallel to nonparallel vectors a and b is
x p = sa + tb
That is, a vector x lies on P exactly when x diers from p by a linear combination of a and b.
5 These

coefficients are also called the weights of the linear combination.

42

1.8.1

CHAPTER 1. VECTORS

Standard Basis Vectors

Definition 1.8.3. For each positive integer n and each j = 1, . . . , n, the j-th standard basis
def
vector in Fn is the vector ej = (ej,1 , . . . , ej,n ) such that

1 if k = j
ej,k =
0 if k 6= j
Less formally, ej is the vector whose j-th component is 1 while all other components are zero.
That is,
2 3
0
6 .. 7
6 . 7
6 7
6 0 7
7
def 6
7
ej = 6
j-th component
6 1 7
6 0 7
6 7
6 . 7
4 .. 5
0

Example 1.8.4. The standard basis vectors in F2 are

1
0
e1 =
and e2 =
0
1

Applied scientists frequently denote the standard basis vectors e1 and e2 in R2 by i and j, respectively.
Example 1.8.5. The standard basis vectors in F3 are
2 3
2 3
1
0
e1 = 4 0 5 , e2 = 4 1 5 and
0
0

3
0
e3 = 4 0 5
1

Applied scientists often denote the standard basis vectors e1 , e2 and e3 in R3 by i, j and k,
respectively.
A very important property of the standard basis vectors is that every vector in Fn is a linear
combination of the standard basis vectors.
2
3
1/2
Example 1.8.6. Express 4 3 5 as a linear combination of the standard basis vectors in R3 .

Solution.

2
4

3
2
1/2
3 5 = 4

3 2 3 2
3
1/2
0
0
0 5+4 3 5+4 0 5
0
0

2 3
2 3
2 3
1
0
0
= ( 1/2) 4 0 5 + 3 4 1 5 + 4 0 5
0
0
1
= ( 1/2)e1 + 3e2 + e3

43

1.8. LINEAR COMBINATIONS

Theorem 1.8.7. Every vector in Fn is a linear combination of the standard basis vectors in Fn .
def

Proof. Let a = (a1 , . . . , an ) 2 Fn . Then

a=

n
X

aj ej

j=1

Therefore, every vector in Fn is a linear combination of the standard basis vectors in Fn .

1.8.2

Parallelotopes

Parallelotopes generalize line segments and parallelograms.

Definition 1.8.8. Let k and n be positive integers. The k-parallelotope in Rn with vertex
p 2 Rn and edges v1 , . . . , vk 2 Rn is
8
9
k
<
=
X
def
P = x 2 Rn : x p =
cj vj for some c1 , . . . , ck 2 [0, 1]
:
;
j=1

Example 1.8.9. A 1-parallelotope is a line segment.

Example 1.8.10. A 2-parallelotope is a parallelogram.

Definition 1.8.11. A parallelepiped is a 3-parallelotope.
See Figure 1.33

1.8.3

Centers of Mass

Definition 1.8.12. The total mass of a system of particles with masses m1 , . . . , mk is

def

M =

k
X

mj

j=1

Definition 1.8.13. The center of mass of a system of particles with masses m1 , . . . , mk and
respective positions r1 , . . . , rk 2 Rn is
def

r =

k
1 X
mj rj
M j=1

By writing
def

r =

k
X
mj
j=1

rj

we see that the center of mass of a system of particles is the linear combination of the particles
positions each weighted by the ratio of its mass to the total mass.
Example 1.8.14.

44

CHAPTER 1. VECTORS
z

v3

v2
v1

Figure 1.33: A Parallelepiped in R3

Find the center of mass of the system of particles described in Table 1.1.
Solution. The total mass of the system is
M = 3 + 8 + 4 = 15
The center of mass of the system is
r=

1
(3(0, 0) + 8(1, 2) + 4(2, 1)) = (16/15, 4/3)
15
N

Terms
linear combination
standard basis vectors
parallelotope

45

1.8. LINEAR COMBINATIONS

Position
(0, 0)
(1, 2)
(2, 1)

Mass (kg)
3
8
4

Table 1.1: The System of Particles in Example 1.8.14

parallelepiped
total mass of a system of particles
center of mass of a system of particles
Concepts
Linear algebra is the study of linear combinations.
Every vector is a linear combination of the standard basis vectors.
Parallelotopes are generalizations of line segments and parallelograms.
The center of mass of a system of particles is the linear combination of the particles positions
each weighted by the ratio of its mass to the total mass.
Skills
Express a vector in Fn as a linear combination of the standard basis vectors in Fn .
Find the center of mass of a system of particles.
Exercises
1. Express ( 2, 0, , 11/17) as a linear combination of the standard basis vectors in R4 .
2. Find the center of mass of the system of particles described in Table 1.2.
Position
(5, 4, 3)
(4, 3, 2)
( 4, 3, 1)
( 9, 8, 6)

Mass (g)
2
5
2
1

Table 1.2: The System of Particles in Exercise 2

46

CHAPTER 1. VECTORS
3. The mass of the Earth is approximately 5.972 1024 kg. The mass of the moon is approximately 7.34767309 1022 kg. The distance from the Earth to the moon is approximately
384, 400 km. How far is the center of mass of the Earth-moon system from the center of the
Earth?
4. In the ammonia molecule (NH3 ), the three hypdrogen (H) atoms form an equilateral triangle,
the distance between centers of the atoms being 16.28 10 11 m, so that the center of the
triangle is 9.40 10 11 m from each hydrogen atom. The nitrogen (N) atom is at the apex
of a pyramid, the three hydrogens constituting the base. The nitrogen-hydrogen distance is
10.14 10 11 m, and the nitrogen-hydrogen mass ratio is 13 : 9. Locate the center of mass of
the ammonia molecule relative to the nitrogen atom. See Figure 1.34.

Figure 1.34: An ammonia molecule

5. Prove that the ratio of the distances of two particles from their center of mass is the inverse
ratio of their masses.

Chapter 2

The Inner Product

In this chapter we define the inner product of two vectors in Fn and use the inner product to develop
the geometric concepts of length, distance and angle in Fn .

2.1

The Inner Product: Motivation

Objective
Motivate the definition of the inner product of two vectors.
def

def

In Section 2.2 we will define the inner product of two vectors a = (a1 , . . . , an ) and b =
(b1 , . . . , bn ) in Fn by
n
X
def
ha, bi =
ak bk
(2.1)
k=1

The inner product will provide a foundation on which we will build the concepts of length, distance
and angle. In this section our goal is provide intuition and motivation for formula (2.1). Like
Section 1.1 this section is informal. None of the definitions or results in this section are official.
How can we determine whether two vectors a and b are perpendicular? That is, under what
conditions do a and b form a right angle? We will solve this problem in the plane using common
notions from elementary geometry and rectangular coordinates. We begin with a familiar fact. The
Pythagorean Theorem states that a triangle with side lengths a b c is a right triangle if and
only if
a2 + b2 = c2
See Figure 2.1.
First, we will use the Pythagorean Theorem to find the distance between two points P and
Q in a plane endowed with a rectangular coordinate system. Denote this distance by |P Q|. Let
(p1 , p2 ) be the coordinates of P , and let (q1 , q2 ) be the coordinates of Q. Consider the point R with
coordinates (q1 , p2 ). Since P and R have the same second coordinate, |P R| = |p1 q1 |. Similarly,
47

48

CHAPTER 2. THE INNER PRODUCT

Figure 2.1: The Pythagorean Theorem

Q (q1 , q2 )

|P Q|

P (p1 , p2 )

|RQ| = |p2

|P R| = |p1

q2 |

R (q1 , p2 )

q1 |

Figure 2.2: Distance Between P and Q

since R and Q have the same first coordinate, |RQ| = |p2
Theorem,
|P Q|2

q2 |. See Figure 2.2. By the Pythagorean

= |P R|2 + |RQ|2
= |p1

q1 |2 + |p2

q2 |2

Thus, the distance between the points P and Q in the plane is

p
|P Q| = (p1 q1 )2 + (p2 q2 )2
kak

|a2 |

|a1 |

Figure 2.3: The Length of a Vector

Second,
we will use the distance formula to define the length of a vector. In particular, if

a1
a =
, then define the length of a by
a2
def

def

kak =

q
a21 + a22

2.2. THE INNER PRODUCT: DEFINITION AND PROPERTIES

49

a+b
b

Figure 2.4: The Pythagorean Theorem

See Figure 2.3.
Finally, recall from the triangle method of adding vectors that two vectors a and b and their
sum a + b form a triangle. By the Pythagorean Theorem, the vectors a and b are perpendicular if
and only if
kak2 + kbk2
q
2 q
2
a21 + a22 +
b21 + b22
a21 + a22 + b21 + b22
a21

a22

b21

b22

= ka + bk2
2
p
=
(a1 + b1 )2 + (a2 + b2 )2
= (a1 + b1 )2 + (a2 + b2 )2

= a21 + 2a1 b1 + b21 + a22 + 2a2 b2 + b22

2(a1 b1 + a2 b2 ) = 0
a1 b1 + a2 b2

= 0

We call the scalar a1 b1 + a2 b2 the inner product of the vectors a and b.

In this section we have used several distinct, yet related, geometric ideas: distance, length, angle
and perpendicularity. In the rest of this chapter we will rigorously develop the concepts of length,
distance and perpendicularity, but starting with the inner product. One advantage of this approach
is that we will have a geometry that works not only in R2 , but also in Fn !
Concept
def

def

Vectors a = (a1 , a2 ) and b = (b1 , b2 ) in R2 are perpendicular if and only if a1 b1 + a2 b2 = 0.

2.2

The Inner Product: Definition and Properties

Objective
Define the inner product in Fn and establish its elementary properties.

50

CHAPTER 2. THE INNER PRODUCT

def

def

Definition 2.2.1. The inner product1 of the vectors a = (a1 , . . . , an ) and b = (b1 , . . . , bn ) in
Fn is
n
X
def
ha, bi =
aj bj
j=1

The inner product of two vectors is a scalar rather than a vector.

def

def

Example 2.2.2. Find the inner product of a = (1

2i, 3i) and b = ( 4 + 5i, 6) in C2 .

Solution. The inner product of a and b is

ha, bi = (1

2i)( 4 + 5i) + (3i)( 6) = (1

2i)( 4

5i)

18i =

14

15i
N

def

def

If a = (a1 , . . . , an ) and b = (b1 , . . . , bn ) are vectors in Rn , then

ha, bi =

n
X

aj bj =

j=1

n
X

aj bj

j=1

because every real number is equal to its complex conjugate.

def

def

Example 2.2.3. Find the inner product of a = (1, 2, 3) and b = ( 4, 5, 6) in R3 .

Solution. The inner product of a and b is
ha, bi = 1( 4) + ( 2)5 + 3( 6) =

32
N

Example 2.2.4. Find the inner product of the standard vectors ej and ek in Fn .
Solution. Denote the components of ej by ej,1 , . . . , ej,n , and denote the components of ek by
ek,1 , . . . , ek,n . Then

1 if l = j
ej,l =
0 if l 6= j
and

ek,l =
By the definition of the inner product,

1
0

hej , ek i =

if l = k
if l =
6 k
n
X

ej,l ek,l

l=1

1 An alternative name for the inner product is the scalar product because the inner product of two vectors is a
scalar rather than a vector. When the vectors have real components we often call the inner product the dot product
because of the traditional notation a b for the inner product of a and b. When the vectors have complex components
we often call the inner product the Hermitian inner product, in honor of Charles Hermite (18221901).

2.2. THE INNER PRODUCT: DEFINITION AND PROPERTIES

If j = k, then
hej , ek i = hej , ej i =

n
X

ej,l ej,l =

l=1

n
X
l=1

51

|ej,l |2 = 1

If j 6= k, then for each l = 1, . . . , n, either ej,l = 0 or ek,l = 0. Thus,

hej , ek i = 0
Therefore,
hej , ek i =

1
0

if j = k
if j =
6 k

The next theorem establishes the fundamental properties of the inner product.
Theorem 2.2.5.
(i) (Positivity) For all a 2 Fn , ha, ai is real and nonnegative.
(ii) (Definiteness) For all a 2 Fn , ha, ai = 0 if and only if a = 0.
(iii) (Conjugate Symmetry) For all a, b 2 Fn ,
ha, bi = hb, ai
(iv) (Additivity in the first argument) For all a, b, c 2 Fn ,
ha + b, ci = ha, ci + hb, ci
(v) (Homogeneity in the first argument) For all a, b 2 Fn and all scalars k,
hka, bi = kha, bi
Proof.
def

(i) Let a = (a1 , . . . , an ) 2 Fn . Because the absolute value of a complex number is a real number,
and squares of real numbers are nonnegative,
ha, ai =

n
X

aj aj =

j=1

n
X
j=1

|aj |2

def

(ii) Let a = (a1 , . . . , an ) 2 Fn . From the proof of part (i),

ha, ai =

n
X
j=1

|aj |2

Since a sum of nonnegative terms is zero exactly when each term is zero, ha, ai = 0 if and
only if aj = 0 for all j = 1, . . . , n; that is, a = 0.

52

CHAPTER 2. THE INNER PRODUCT

def

def

(iii) Let a = (a1 , . . . , an ), b = (b1 , . . . , bn ) 2 Fn . By properties of complex conjugates,

ha, bi =
def

n
X

aj bj =

j=1

n
X

aj bj =

j=1

def

n
X

aj bj =

j=1

n
X

bj aj =

j=1

n
X

bj aj = hb, ai

n
X

bj cj = ha, ci + hb, ci

j=1

def

(iv) Let a = (a1 , . . . , an ), b = (b1 , . . . , bn ), c = (c1 , . . . , cn ) 2 Fn .

ha + b, ci =
def

n
X

(aj + bj )cj =

j=1

n
X

(aj cj + bj cj ) =

j=1

n
X

aj cj +

j=1

j=1

def

(v) Let a = (a1 , . . . , an ), b = (b1 , . . . , bn ) 2 Fn . Let k be a scalar.

hka, bi =

n
X
j=1

(kaj )bj =

n
X

k(aj bj ) = k

j=1

n
X

aj bj = kha, bi

j=1

From the additivity and homogeneity of the inner product in its first argument and conjugate
symmetry we derive corresponding properties of the inner product in its second argument.
Corollary 2.2.6.
(i) (Additivity in the second argument) For all a, b, c 2 Fn ,
ha, b + ci = ha, bi + ha, ci
(ii) (Conjugate Homogeneity in the second argument) For all a, b 2 Fn and all scalars k,
ha, kbi = kha, bi
Proof.
(i) Let a, b, c 2 Fn . By conjugate symmetry and the additivity in the first argument,
ha, b + ci = hb + c, ai = hb, ai + hc, ai = hb, ai + hc, ai = ha, bi + ha, ci
(ii) Let a, b 2 Fn . Let k be a scalar. By conjugate symmetry and the homogeneity in the first
argument,
ha, kbi = hkb, ai = khb, ai = khb, ai = kha, bi
From additivity and homogeneity in the first argument we obtain linearity in the first argument:
hka + lb, ci = hka, ci + hlb, ci = kha, ci + lhb, ci
for all a, b, c 2 Fn and all scalars k, l. Our next theorem extends linearity in the first argument to
any number of terms. The proof uses the Principle of Mathematical Induction.

53

2.2. THE INNER PRODUCT: DEFINITION AND PROPERTIES

Theorem 2.2.7. For each integer p 2, if a1 , . . . , ap , b 2 Fn and k1 , . . . , kp are scalars, then
* p
+
p
X
X
kj aj , b =
kj haj , bi
j=1

j=1

Proof. We proceed by means of induction on the number p of vectors. If p = 2, then

hk1 a1 + k2 a2 , bi = hk1 a1 , bi + hk2 a2 , bi

= k1 ha1 , bi + k2 ha2 , bi

for all a1 , a2 2 Fn and all scalars k1 and k2 .

Let l
2 be an integer. As an inductive hypothesis, assume that if a1 , . . . , al , b 2 Fn and
k1 , . . . , kl are scalars, then
* l
+
l
X
X
kj aj , b =
kj haj , bi
j=1

j=1

Let a1 , . . . , al+1 , b 2 F , and let k1 , . . . , kl+1 be scalars. From additivity in the first argument, the
inductive hypothesis and homogeneity in the first argument we conclude that
1
* l+1
+
*0 l
+
X
X
@
kj aj , b
=
kj aj A + kl+1 al+1 , b
n

j=1

j=1

1 +
*0 l
X
@
=
kj aj A , b + hkl+1 al+1 , bi
j=1

0
1
l
X
= @
kj haj , biA + kl+1 hal+1 , bi
j=1

l+1
X
j=1

kj haj , bi

By the Principle of Mathematical Induction, for each integer p

k1 , . . . , kp are scalars, then
* p
+
p
X
X
kj aj , b =
kj haj , bi
j=1

2, if a1 , . . . , ap , b 2 Fn and

j=1

From linearity in the first argument and conjugate symmetry we obtain antilinearity in the
second argument:
ha, kb + lci = hkb + lc, ai

= khb, ai + lhc, ai
= khb, ai + lhc, ai
= khb, ai + lhc, ai
= kha, bi + lha, ci

54

CHAPTER 2. THE INNER PRODUCT

for all a, b, c 2 Fn and all scalars k, l. In the exercises we ask you to extend antilinearity in the
second argument to any number of terms. Together, the property of linearity in the first argument
and antilinearity in the second is called sesquilinearity.
The following proposition shows that we can use the inner product to extract individual components from a coordinate vector.
Proposition 2.2.8. For each a = (a1 , . . . , an ) 2 Fn and each k = 1, . . . , n
ha, ek i = ak
def

Proof. For each a = (a1 , . . . , an ) 2 Fn and each k = 1, . . . , n,

* n
+
n
X
X
ha, ek i =
aj ej , ek =
aj hej , ek i = aj
j=1

j=1

by Theorem 2.2.7 and Example 2.2.4.

Terms
inner product
positivity
definiteness
conjugate symmetry
additivity in the first argument
homogeneity in the first argument
additivity in the second argument
conjugate homogeneity in the first argument
linearity in the first argument
antilinearity in the second argument
sesquilinearity
Concept
The inner product is positive, definite, conjugate symmetric, additive in both arguments,
homogeneous in the first argument and conjugate homogeneous in the second argument.
Skill
Find the inner product of two vectors in Fn .

55

2.3. NORM AND DISTANCE

Exercises
1. Find the inner product of each of the following pairs of vectors.
(a) (1

2i, 3 + 4i), ( 2 + i, 4 + 3i)

(b) (1, 5), ( 2, 3)

(c) ( 1, 0, 7), (2, 4, 6)
(d) (1, 3, 5, . . . , 2n

1), (2, 4, 6, . . . , ( 1)n+1 2n)

2. Prove that for each positive integer p, if a, b1 , . . . , bp 2 Fn and k1 , . . . , kp are scalars, then
*
+
p
p
X
X
a,
kj bj =
kj ha, bj i
j=1

j=1

3. (a) Provide an example of vectors a, b, x 2 Fn so that a 6= b, x 6= 0 and hx, ai = hx, bi.

(b) Prove that if hx, ai = 0 for all x 2 Fn , then a = 0.

(c) Prove that if a, b 2 Fn and hx, ai = hx, bi for all x 2 Fn , then a = b.

2.3

Norm and Distance

Objective
Define the norm in Fn , and establish its elementary properties.
Definition 2.3.1. The norm2 of a 2 Fn is
def

kak =
def

(2.2)

ha, ai

If a = (a1 , . . . an ), then by the definition of the inner product

0
1 12
n
X
kak = @
|aj |2 A
j=1

Squaring each side of equations (2.2) and (2.3) we obtain

kak2 = ha, ai =
def

n
X
j=1

Example 2.3.2. Find the norm of a = (1, 2, 3) in R3 .

2 length

or magnitude

|aj |2

(2.3)

56

CHAPTER 2. THE INNER PRODUCT

Solution. The norm of a is

kak = k(1, 2, 3)k =

12 + |

2|2 + 32 =

14
N

Example 2.3.3. Find the norm of (1, i) in C2 .

Solution. The norm of (1, i) is
k(1, i)k =

p
p
12 + |i|2 = 2

If we were to add the squares of the components of (1, i) without conjugating, we would obtain
p
p
12 + i2 = 1 + ( 1) = 0

Thus, there would be a nonzero vector with zero length. This example shows why conjugation is
desirable in the definition of the inner product on Cn .
N
The next theorem establishes some fundamental properties of the norm.
Theorem 2.3.4.
(i) For all a 2 Fn , kak is real and nonnegative.
(ii) For all a 2 Fn , kak = 0 if and only if a = 0.
(iii) For all a 2 Fn and all scalars k,
For every z 2 C,
zz = |z|

kkak = |k|kak

Proof.
(i) For all a 2 Fn , kak =

ha, ai which is real and nonnegative.

(ii) For all a 2 Fn , kak = 0 if and only if ha, ai = 0; that is, a = 0 by the definiteness of the inner
product.
(iii) Let a 2 Fn . Let k be a scalar. By homogeneity in the first argument and conjugate homogeneity in the second we have
q
p
p
kkak = hka, kai = (kk)ha, ai = |k|2 ha, ai = |k|kak
It will be convenient to have a formula for the square of the norm of a sum of two vectors.
Proposition 2.3.5.
for all a, b 2 Fn .

ka + bk2 = kak2 + 2Re ha, bi + kbk2

57

2.3. NORM AND DISTANCE

For every z 2 C,
z + z = 2Re z

Proof. For all a, b 2 Fn ,

ka + bk2

= ha + b, a + bi

= ha, a + bi + hb, a + bi

= ha, ai + ha, bi + hb, ai + hb, bi

= ha, ai + ha, bi + ha, bi + hb, bi
= kak2 + 2Re ha, bi + kbk2

2.3.1

Normalization

Often it is useful to specify a direction in Fn by providing a vector in that direction with length
one.
Definition 2.3.6. A vector u 2 Fn is a unit vector if kuk = 1.
The next theorem shows how to find a unique unit vector in the direction of any nonzero vector.
To normalize a nonzero vector is to find a unit vector in the same direction.
Theorem 2.3.7. If a is a nonzero vector in Fn , then
of a.
Proof. Since a 6= 0 we have

1
a is the unique unit vector in the direction
kak

1
1
> 0. So,
a is a vector in the direction of a. By Theorem 2.3.4,
kak
kak
1
1
kak
a =
kak =
=1
kak
kak
kak

1
def 1
a is a unit vector in the direction of a. Define u =
a.
kak
kak
To prove that u is unique, assume that v is a unit vector in the direction of a. Then kvk = 1
and there exists a nonnegative scalar k such that
Therefore,

v = ka
Consequently,
kkak

|k|kak

= 1
= 1

|k| = 1/kak

Since k

0, we obtain k = 1/kak. Hence,

v=

1
a=u
kak

Therefore, u is the unique unit vector in the direction of a.

58

CHAPTER 2. THE INNER PRODUCT

For nonzero vectors a we denote

1
a
a by
.
kak
kak

def

Example 2.3.8. Normalize a = (1, 2, 3).

Solution. Applying Theorem 2.3.7, a unit vector in the direction of a is

a
(1, 2, 3)
(1, 2, 3)
1
2
3
u=
=
= p
= p , p ,p
kak
k(1, 2, 3)k
14
14
14
14

2.3.2

Polarization

While we define the norm of a vector a 2 Fn in terms of the inner product by the formula
def

kak =

p
ha, ai

there is also an expression for the inner product in terms of the norm.
Polarization Identity.
ha, bi =

1
ka + bk2
4

ka

bk2

for all a, b 2 Rn .
Proof. Let a, b 2 Rn . By Proposition 2.3.5 and Corollary 2.2.6(ii),
1
ka + bk2
4

ka

bk2

1
kak2 + 2ha, bi + kbk2
4
1
=
(2ha, bi + 2ha, bi)
4
= ha, bi
=

kak2

2ha, bi

bk2

A generalized version of the polarization identity for vectors with complex components appears
in the exercises.

Distance

kb

ak

2.3.3

Figure 2.5: The Distance Between Vectors

59

2.3. NORM AND DISTANCE

Definition 2.3.9. The distance between vectors a and b in Fn is
def

dist(a, b) = kb
def

ak

def

If a = (a1 , . . . , an ) and b = (b1 , . . . , bn ), then

0
n
X
dist(a, b) = @
|bj
j=1
def

1 12

aj |2 A

def

Example 2.3.10. Find the distance between a = (1, 2, 3) and b = ( 4, 5, 6).

Solution. The distance between a and b is
p
dist(a, b) = ( 4 1)2 + (5

( 2))2 + ( 6

3)2 =

155
N

Terms
norm
unit vector
normalize
distance
Concepts
For each nonzero vector a in Fn there is a unique unit vector in the direction of a.
The Generalized Polarization Identity expresses the inner product of two vectors in terms of
norms.
Skills
Find the norm of a vector in Fn .
Normalize a nonzero vector in Fn .
Find the distance between two vectors in Fn .
Exercises
1. Find the norm of each vector.
(a) (1

2i, 3 + 4i)

60

CHAPTER 2. THE INNER PRODUCT

(b)
(c)
(d)
(e)

(1, 5)
( 1, 0, 7)
(1, 3, 5, . . . , 2n 1),
(2, 4, 6, . . . , ( 1)n+1 2n)

2. Normalize each vector.

(a) (8/3, 1)
(b) (7/4, 1/2, 1)
3. Find a vector with norm 3 in the direction of the vector (1, 1, 1).
4. Find the distance between each pair of vectors.
(a) (1 2i, 3 + 4i), ( 2 + i, 4 + 3i)
(b) (10, 3), ( 1, 5)
(c) (0, 5, 2), ( 4, 1, 8)
5. Prove that for all a, b 2 Rn , kak = kbk if and only if ha + b, a
6. Prove that

kak
def

for all a = (a1 , . . . , an ) 2 Fn .

n
X
i=1

|ai |

bi = 0.

nkak

7. (Parallelogram Equality) Prove that

ka + bk2 + ka

bk2 = 2 kak2 + kbk2

for all a, b 2 Fn . Draw a diagram of vectors a, b, a + b and a

what the parallelogram equality says about parallelograms.
8. Prove that

ha + b, a

for all a, b 2 Fn .

bi = ka||2

2iIm ha, bi

b. Use the diagram to explain

kbk2

9. Prove that

(a) dist(a, b) 0 with equality if and only if a = b

(b) dist(a, b) = dist(b, a)
for all a, b 2 Fn .

10. (a) (Generalized Polarization Identity) Prove that

3

ha, bi =

1X k
i ka + ik bk2
4
k=0

for all a, b 2 C .
(b) Show that the Polarization Identity is a consequence of the Generalized Polarization
Identity.
n

61

2.4. ORTHOGONALITY

2.4

Orthogonality

Objective
Define orthogonality.
Definition 2.4.1. Vectors a and b in Fn are orthogonal3 , denoted a ? b, if
ha, bi = 0
The zero vector is orthogonal to every vector a 2 Fn since ha, 0i = 0.

x
y
Example 2.4.2. The vectors
and
in R2 are orthogonal because
y
x

x
y

y
x

= x( y) + yx = 0

If we regard vectors a and b as the sides of a triangle, then the third side of the triangle is
a + b. Our definition of orthogonality agrees with our intuition to the extent that the Pythagorean
Theorem holds in Rn (although not in Cn ) exactly when ha, bi = 0. See Figure 2.6.

a+b
b

Figure 2.6: The Pythagorean Theorem

Pythagorean Theorem. If a, b 2 Fn are orthogonal, then
ka + bk2 = kak2 + kbk2
The converse is true if F = R.
Proof. Let a, b 2 Fn . By Proposition 2.3.5
ka + bk2 = kak2 + 2Re ha, bi + kbk2
3 The term orthogonal comes from the Greek orthogonios for right-angled. An alternative term is perpendicular
which comes from the Latin perpendicularis for vertical.

62

CHAPTER 2. THE INNER PRODUCT

If a and b are orthogonal, then ha, bi = 0. Thus,

ka + bk2 = kak2 + kbk2
Conversely, if F = R and ka + bk2 = kak2 + kbk2 , then Re ha, bi = ha, bi. Hence,
kak2 + 2Re ha, bi + kbk2

= kak2 + kbk2

ha, bi = 0

Therefore, a and b are orthogonal.

Proposition 2.4.3. Let a 2 Fn . If b, c 2 Fn are orthogonal to a, then b + c is orthogonal to a.
Proof. Assume that b, c 2 Fn so that b and c are orthogonal to a. By additivity of the inner
product in the first argument and the definition of orthogonality,
hb + c, ai = hb, ai + hc, ai = 0
Therefore, b + c is orthogonal to a.
Term
orthogonal
Concept
Pythagorean Theorem
Skill
Determine whether two vectors in Fn are orthogonal.
Exercises
1. Determine whether each of the following statements is true or false. If the statement is true,
then prove it. If the statement is false, then provide a counterexample. Let a, b and c be
vectors in Fn .
(a) a ? a

(b) If a ? b, then b ? a.

(c) If a ? b and b ? c, then a ? c.

2. Determine which pairs of vectors are orthogonal.

(a) (1

2i, 3 + 4i), ( 2 + i, 4 + 3i)

(b) (8, 5), ( 2, 3)

(c) (12, 3, 5), (2, 3, 3)

2.5. ORTHOGONAL DECOMPOSITION AND ORTHOGONAL PROJECTION

63

(d) (3, 2, 5, 0), ( 4, 1, 2, 6)

3. Find a real number k so that (1, 2, k, 3) and (3, k, 7, 5) are orthogonal in R4 .
4. Provide an example of vectors a and b in C2 so that ka + bk2 = kak2 + kbk2 , but a and b
are not orthogonal.
5. A rectangle is a parallelogram with orthogonal edges. Prove that the diagonals of a parallelogram in Rn have the same norm if and only if the parallelogram is a rectangle.
6. A rhombus is a parallelogram with edges of equal norm. Prove that the diagonals of a
parallelogram in Rn are orthogonal if and only if the parallelogram is a rhombus.

2.5

Orthogonal Decomposition and Orthogonal Projection

Objective
Develop the orthogonal decomposition of one vector with respect to another and the orthogonal projection of one vector onto another.
Given a vector a 2 Fn our goal is to express each vector b 2 Fn as a sum
b = bk + b?

where bk is parallel to a and b? is orthogonal to a. See Figure 2.7. This sum is the orthogonal
decomposition of b with respect to a. We will see that this decomposition is unique whenever a is
nonzero.
b

b?

bk

Figure 2.7: The Orthogonal Decomposition

Theorem 2.5.1. Let a 2 Fn . For each b 2 Fn there are vectors bk and b? in Fn such that bk is
parallel to a, b? is orthogonal to a, and
b = bk + b?
In particular, if a 6= 0, then bk and b? are unique, and
bk = hb, uiu

where u is a unit vector in the direction of a.

64

CHAPTER 2. THE INNER PRODUCT

Proof. Let a, b 2 Fn . If a = 0, then let bk = 0 and b? = b.

If a 6= 0, then u = a/kak is a unit vector in the direction of a. Define
def

bk = hb, uiu
Then
bk =
so that bk is parallel to a. Define

hb, ui
a
kak

def

b? = b
Adding bk to both sides we obtain

bk

b = bk + b?

We will show that b? is orthogonal to a.

hb? , ai = hb

bk , ai

= hb, ai

hbk , ai

= hb, ai

hb, uihu, ai

= hb, ai
= hb, ai
= hb, ai
= 0

hhb, uiu, ai
hb, (1/kak)aih(1/kak)a, ai
ha, ai
hb, ai
kak2

Hence, b? is orthogonal to a.
To prove that bk and b? are unique, assume that ck and c? are vectors in Fn so that ck is
parallel to a, c? is orthogonal to a, and b = ck + c? . Since ck is parallel to a 6= 0, there is a scalar
k so that ck = ka. Then b = ck + c? implies that
c? = b

ka

Because c? is orthogonal to a we obtain

0

= hc? , ai
= hb

ka, ai

= hb, ai
= hb, ai

kha, ai
kkak2

Solving for k we obtain

k

=
=
=

hb, ai
kak2
hb, a/kaki
kak
hb, ui
kak

2.5. ORTHOGONAL DECOMPOSITION AND ORTHOGONAL PROJECTION

65

proja b

Figure 2.8: The Orthogonal Projection

Thus,
ck = ka =
and

hb, ui
a = hb, uiu = bk
kak

c? = b

ck = b

bk = b?

Therefore, there are unique vectors bk and b? in Fn so that bk is parallel to a, b? is orthogonal

to a, and b = bk + b? .
Theorem 2.5.1 motivates the following definition. See Figure 2.8.
Definition 2.5.2. Let a, b 2 Fn . If a 6= 0, then the orthogonal projection of b onto a is
def

proja b = hb, ui u
where u is a unit vector in the direction of a. If a = 0, then the projection of b onto a is the zero
vector.
def

def

Example 2.5.3. Let a = (1, 2, 3) and b = ( 4, 5, 6). Find proja b.

Solution 1. Normalize a to obtain
u=
Then
proja b = hb, ui u =

a
=
kak

1
p ,
14

32
1
p
p ,
14
14

2
3
p ,p
14
14

2
3
p ,p
14
14

16 32
, ,
7 7

48
7

Solution 2. Often it is easier to compute a projection without explicitly normalizing a.

hb, ai
32
16 32 48
proja b =
a
=
(1,
2,
3)
=
,
,
kak2
14
7 7
7

66

CHAPTER 2. THE INNER PRODUCT

Term
orthogonal projection
Concepts
Orthogonal Decomposition: Given a vector a 2 Fn each vector b 2 Fn is a sum
b = bk + b?
where bk is parallel to a and b? is orthogonal to a. This decomposition is unique when a is
nonzero.

Skills
Find the orthogonal decomposition of one vector in Fn with respect to another.
Find the orthogonal projection of one vector in Fn onto another.
Exercises
1. Express b as the sum of a vector parallel to a and a vector orthogonal to a.
def

def

(a) a = (4, 7), b = (2, 3)

def

def

(b) a = (7, 1), b = (2, 6)

2. Find proja b.
p
p
def
def
(a) a = (2, 3, 1), b = (1/ 2, 1/ 2, 0)
def

def

(b) a = (1, 1, 7, 3, 2), b = (2, 5, 0, 9, 1)

def

def

(c) a = (5 + i, 2i), b = (3 + 4i, 2

def

(d) a = (1

def

i, 3i, 1 + i), b = (1, 2

3i)
i, 3 + 2i)

3. Let a, b 2 Fn .
(a) Prove that proja b and b
(b) Prove that b

proja b are orthogonal.

proja b = 0 if and only if a is parallel to b.

67

2.6. THE CAUCHY-SCHWARZ INEQUALITY

2.6

The Cauchy-Schwarz Inequality

Objective
State and prove the Cauchy-Schwarz inequality and explain its major consequences.
From the orthogonal decomposition and the Pythagorean Theorem we obtain one of the most
important facts in mathematics: the Cauchy-Schwarz inequality.
Cauchy-Schwarz Inequality. For all a, b 2 Fn
| ha, bi | kak kbk
with equality if and only if a is parallel to b.
Proof. Let a, b 2 Fn . If a = 0, then both sides of the inequality are zero. So, equality holds in this
case, and a k b.
If a 6= 0, then u = a/kak is a unit vector in the direction of a. Since the vectors proja b and
b proja b are orthogonal, the Pythagorean Theorem implies that
kbk2

= kproja bk2 + kb

proja bk2

khb, uiuk2

| hb, ui |2 kuk2

2
a
b,
kak
| hb, ai |2
kak2
with equality if and only if b
we have

proja b = 0, i.e. a is parallel to b. Multiplying each side by kak2

kak2 kbk2

| hb, ai |2

Taking square roots and applying the fact that | ha, bi | = | hb, ai | we obtain
| ha, bi | kak kbk
with equality if and only if a is parallel to b.

2.6.1

The Triangle Inequality

From the Cauchy-Schwarz inequality we derive another fundamental property of the norm called
the triangle inequality. Remember that if we regard vectors a and b as sides of a triangle, then the
third side of the triangle is a + b. Geometrically, the triangle inequality says that the length of one
side of a triangle is no greater than the sum of the lengths of the other two sides. See Figure 2.9.

68

CHAPTER 2. THE INNER PRODUCT

a+b

Figure 2.9: The Triangle Inequality

Triangle Inequality. For all a, b 2 Fn
ka + bk kak + kbk
with equality if and only if a is in the direction of b.
Proof. Let a, b 2 Fn . By Proposition 2.3.5 and the Cauchy-Schwarz inequality,
ka + bk2

= kak2 + 2Re ha, bi + kbk2

kak2 + 2|Re ha, bi | + kbk2

kak2 + 2| ha, bi | + kbk2
kak2 + 2kak kbk + kbk2
(kak + kbk)2

Taking square roots we have

ka + bk kak + kbk
Equality holds if and only if
Re ha, bi = | ha, bi | = kak kbk
By Cauchys inequality, | ha, bi | = kak kbk if and only if a and b are parallel. That is, either there
is a scalar k so that b = ka or a = 0. If there is a scalar k so that b = ka, then Re ha, bi = | ha, bi |
if and only if
Re kkak2 = |k| kak2
So, either k 0 or a = 0. In either case, a is in the direction of b. Therefore, equality holds if and
only if a is in the direction of b.

2.6.2
In this subsection all

Angle

If a and b are nonzero vectors in Rn , then the Cauchy-Schwarz inequality implies that

vectors have real

entries.

| ha, bi |
1
kak kbk

69

2.6. THE CAUCHY-SCHWARZ INEQUALITY

b

](a, b)

Figure 2.10: The Angle Between Nonzero Vectors

By the definition of absolute value,
1
The arccosine of
x 2 [

ha, bi
1
kak kbk

Since the domain of the arccosine function is [ 1, 1],

1, 1] is the real

number 2 [0, ] such

arccos

that cos = x.

ha, bi
kak kbk

is defined.
Definition 2.6.1. The angle between nonzero vectors a and b in Rn is
def

](a, b) = arccos

ha, bi
kak kbk

If a = 0 or b = 0, then we do not define the angle between a and b. By the definition of the
arccosine function, the angle between two nonzero vectors is a number between 0 and .
def

def

Example 2.6.2. Find the angle between a = (1, 2, 3) and b = ( 4, 5, 6).

def

def

Solution. The angle between a = (1, 2, 3) and b = ( 4, 5, 6) is

ha, bi
32
32
p p
p
](a, b) = arccos
= arccos
= arccos
2.9 rad
kak kbk
14 77
7 22
N
Equipped with this definition of angle we see that our definition of orthogonality (Definition 2.4.1) agrees with our intuition. Indeed, nonzero vectors a and b are orthogonal if and only
if
ha, bi
0

](a, b) = arccos
= arccos
= arccos 0 = rad
kak kbk
kak kbk
2
Theorem 2.6.3. If is the angle between nonzero vectors a, b 2 Rn , then
ha, bi = kak kbk cos

70

CHAPTER 2. THE INNER PRODUCT

Proof. Let a, b 2 Rn be nonzero. Let be the angle between a and b. So,

cos =

ha, bi
kak kbk

Multiplying both sides by kak kbk we obtain

ha, bi = kak kbk cos .
If we abide by the convention that kak kbk cos = 0 whenever a = 0 or b = 0, then
ha, bi = kak kbk cos
for all vectors a, b 2 Rn .
Theorem 2.6.4. If is the angle between the vectors a and b in Rn , then
kb? k = kbk sin
where b? is the component of b orthogonal to a.
Proof. By the Pythagorean Theorem,
kb? k2

= kb

proja (b)k2

= kbk

= kbk2
= kbk2
= kbk2 (1

Since 0 , we know that sin

kproja (b)k2
| ha, bi |2
kak2
kak4
kak2 kbk2 cos2
kak2
cos2 )

= kbk2 sin2
p
0. So, sin2 = | sin | = sin . Therefore,
kb? k = kbk sin

2.6.3
In this subsection all
vectors have real
entries.

Work

Recall that the work W done by a constant scalar force F acting on a particle moving through a
displacement d along a line is the product of the force and the displacement:
W = Fd
Now consider a constant nonzero vector force F 2 Rn acting a particle moving through a nonzero
displacement D 2 Rn . The force F itself may or may not be displacing the particle. On the one

71

2.6. THE CAUCHY-SCHWARZ INEQUALITY

F

}
|W|

Figure 2.11: Work

hand, the force may be moving a wagon as in Figure 2.11. On the other hand, the force of a
constant wind blowing against you as you walk down a sidewalk acts on you, but hopefully does
not displace you.
By analogy with the scalar case, the work W done by the force F on the particle is the product
of the component of the force parallel to the displacement and the magnitude of the displacement.
Orthogonally decomposing the force F with respect to the displacement D we obtain
F = F? + Fk
= F? + hF, uiu
where u is the unit vector in the direction of D. The component of the force parallel to D is

D
hF, Di
hF, ui = F,
=
= kFk cos
kDk
kDk
where is the angle between F and D. Therefore,

W = kFk kDk cos

Indeed, we define work using this formula.
Definition 2.6.5. The work done by a constant force F 2 Rn on a particle with displacement
D 2 Rn is
def
W = kFk kDk cos
where is the angle between F and D.

The SI units for the magnitudes of force and displacement are the newton (N) and the meter
(m), respectively. The SI unit for work is the joule (J) where 1 J = 1 N m.
Example 2.6.6. A 12 N force acts on a box at angle of /9 radians. Find the work done by the
force on the box as the box moves a distance of 3 m.
Solution. The work is

W = kFk kdk cos = (12 N)(3 m) cos

33.8 J
9

72

CHAPTER 2. THE INNER PRODUCT

Theorem 2.6.7. The work W done by a constant force F 2 Rn on an object with displacement
D 2 Rn is
W = hF, Di
Proof. If is the angle between F and D, then
W = kFk kDk cos = hF, Di
by Theorem 2.6.3.
Terms
angle
work
Concepts
Cauchy-Schwarz inequality
Triangle inequality
If is the angle between nonzero vectors a and b in Rn , then
ha, bi = kakkbk cos
The work W done by a constant force F 2 Rn on an object with displacement D 2 Rn is
W = hF, Di
Skills
Find the angle between two nonzero vectors in Rn .
Find the work done by a constant force F on a particle displaced by D.
Exercises
1. Let n be a positive integer. Verify the Cauchy-Schwarz inequality for (1, 2, 3, . . . , n) and
(1, 1, 1, . . . , 1) in Rn .
2. Let n be a positive even integer. Verify the triangle inequality for (1, 0, 1, 0, . . . , 0) and
(0, 1, 0, 1, . . . , 1) in Rn .
3. (Generalized Triangle Inequality) Prove that for each positive integer p
then
p
p
X
X
ai
kai k
i=1

i=1

2, if a1 , . . . , ap 2 Fn ,

73

2.6. THE CAUCHY-SCHWARZ INEQUALITY

4. Prove that
dist(a, c) dist(a, b) + dist(b, c)

for all a, b, c 2 F .
n

5. (Reverse Triangle Inequality) Prove that

|kak

bk

kbk| ka

for all a, b 2 Fn .
6. Find the angle between each of the following pairs of vectors.
p
p
(a) ( 3, 1), (
3, 1)
(b) (1, 2, 3), (3, 6, 5)
7. Prove that if ka

bk > ka + bk, then ](a, b) > /2 for all nonzero a, b 2 Rn .

8. (Law of Cosines) Prove that

ka

bk2 = kak2 + kbk2

2kak kbk cos

for all a, b 2 Rn where is the angle between a and b.

9. Prove that for all nonzero a, b 2 Rn , the vector kbka + kakb bisects the angle between a and
b.
def

10. The direction cosines of a nonzero vector a = (a1 , a2 , a3 ) 2 R3 are the cosines of the angles
, and between a and (1, 0, 0), (0, 1, 0) and (0, 0, 1), respectively. Express the direction
cosines cos , cos and cos in terms of a1 , a2 and a3 .
11. A student lifts a 2 kg textbook (with negligible acceleration) to a height of 1.8 m to place it
on a bookshelf.
(a) How much work does the student do on the textbook?
(b) How much work does the force of gravity do on the textbook?
(c) What is the total work done by both of the forces acting on the textbook?
12. A math professor pushes a filing cabinet whose mass is 85 kg at a constant speed across a
carpeted floor for a distance of 3.1 m. The carpet exerts a frictional force on the cabinet
proportional to cabinets weight with constant of proportionality 0.22.
(a) How much work does the professor do on the cabinet?
(b) How much work does the frictional force do on the cabinet?
(c) How much work does gravity do on the cabinet?
(d) How much work does the carpet do on the cabinet (excluding the frictional force)?
13. Find the work done in pushing a handtruck loaded with 230 kg of math textbooks 13 m up
a ramp inclined 30 from horizontal.

74

CHAPTER 2. THE INNER PRODUCT

14. A winch pulls a crate with mass 15 kg a distance 5.7 m up a frictionless ramp to a height 2.5
m above its starting point.
(a) What force does the winch exert on the crate?
(b) How much work is done by the winch on the crate?
(c) How much work is required to lift the crate vertically upward through a height of 2.5 m?

Chapter 3

Systems of Linear Equations

3.1

Systems of Linear Equations

Objectives
Introduce systems of linear equations and their solution sets.
Solve systems of linear equations by applying two operations to the equations in the system:
multiplying both sides of an equation by a nonzero scalar, and/or
adding or subtracting the left- and right-hand sides of one equation to/from the left- and
right-hand sides, respectively, of another equation.
Introduce the fundamental questions of consistency and uniqueness of solutions.
In Sections 1.6 and 1.7 we encountered standard linear equations in two and three variables,
respectively. Now we extend our consideration of linear equations to more than three variables.
Definition 3.1.1. A linear equation is an equation of the form
a1 x1 + a2 x2 + a3 x3 + + an xn = b
where n is a positive integer and a1 , . . . , an and b are scalars.
Definition 3.1.2. Let m and n be positive integers. A system of m linear equations in n
variables (or an m n linear system) is a collection of m linear equations
8
a1,1 x1 + + a1,j xj + + a1,n xn = b1
>
>
>
>
..
>
>
.
<
ai,1 x1 + + ai,j xj + + ai,n xn = bi
>
>
..
>
>
>
.
>
:
am,1 x1 + + am,j xj + + am,n xn = bm
75

76

CHAPTER 3. SYSTEMS OF LINEAR EQUATIONS

where bi are scalars for i = 1, . . . , m, and ai,j are scalars for i = 1, . . . , m and j = 1, . . . , n. The
scalars ai,j are called the coefficients of the system.
Definition 3.1.3. A solution of a linear system
8
a1,1 x1 + + a1,j xj + + a1,n xn
>
>
>
>
>
>
<
ai,1 x1 + + ai,j xj + + ai,n xn
>
>
>
>
>
>
:
am,1 x1 + + am,j xj + + am,n xn

=
..
.
=
..
.

b1
bi

= bm

is a vector (x1 , . . . , xn ) in Rn or Cn such that the components x1 , . . . , xn satisfy each equation in

the system.
Definition 3.1.4. The solution set of a linear system is the set of all solutions of the system.
To solve a linear system is to find the systems solution set.
Example 3.1.5. Let a, b and c be real numbers so that at least one of a and b is nonzero. The
solution set of the 1 2 linear system
ax + by = c
is a line in R2 .
Example 3.1.6. Let a, b, c and d be real numbers so that at least one of a, b and c is nonzero. The
solution set of the 1 3 linear system
ax + by + cz = d
is a plane in R3 .
Our goal is to solve linear systems.
Definition 3.1.7. Two linear systems are equivalent if their solution sets are equal.
Our strategy for solving a linear system will be to transform the system into an equivalent
system whose solution set is more apparent. To obtain an equivalent system we will apply two
operations to the equations in the system:
Multiply both sides of an equation by a nonzero scalar.
Add or subtract the left- and right-hand sides of one equation in the system to the left- and
right-hand sides, respectively, of another equation in the system.
The following theorem shows that each of these operations produces a system equivalent to the
original. Often we combine these two operations in one step: adding a nonzero multiple of the leftand right-hand sides of one equation in the system to the left- and right-hand sides, respectively,
of another equation in the system.
Theorem 3.1.8. Let ai,j and bi be scalars for all i = 1, . . . , m and j = 1, . . . , n.

77

3.1. SYSTEMS OF LINEAR EQUATIONS

(i) Let c be a nonzero scalar. The linear system
8
a1,1 x1 + + a1,n xn
>
>
>
>
>
>
<
ai,1 x1 + + ai,n xn
>
>
>
>
>
>
:
am,1 x1 + + am,n xn
is equivalent to the linear system
8
a1,1 x1
+
>
>
>
>
>
>
<
c(ai,1 x1 ) +
>
>
>
>
>
>
:
am,1 x1 +

=
..
.

b1

=
..
.

bi

= bm

a1,n xn

=
..
.

b1

bm

+ c(ai,n xn ) = cbi
..
.
+

am,n xn

a1,n xn

ai,n xn

ak,n xn

+ am,n xn

for all i = 1, . . . , m.
(ii) The linear system

8
a1,1 x1
>
>
>
>
>
>
>
>
>
>
>
< ai,1 x1

>
>
>
ak,1 x1
>
>
>
>
>
>
>
>
:
am,1 x1

is equivalent to the linear system

8
a1,1 x1
+
>
>
>
>
>
>
>
>
>
>
ai,1 x1
+
>
<
>
>
>
(ak,1 ai,1 )x1
>
>
>
>
>
>
>
>
:
am,1 x1

a1,n xn

ai,n xn

=
..
.
=
..
.
=
..
.

bi
bk

= bm

+ (ak,n ai,n )xn

am,n xn

b1

=
..
.

b1

=
..
.

bi

bm

= bk bi
..
.

for all i = 1, . . . , m and all k = 1, . . . , m.

Proof.
(i) Let c be a nonzero scalar. Let i 2 {1, . . . , m}. The strategy of the proof is to show that every
solution of the original system is a solution of the transformed system and that every solution
of the transformed system is a solution of the original system.

78

CHAPTER 3. SYSTEMS OF LINEAR EQUATIONS

Let (x1 , . . . , xn ) be a solution of
8
a1,1 x1
>
>
>
>
..
>
>
.
<
ai,1 x1
>
>
..
>
>
>
.
>
:
am,1 x1

+ am,n xn

a1,n xn
..
.
ai,n xn
..
.

b1
..
.

bi
..
.

= bm

For all l 2 {1, . . . , m} such that l 6= k,

al,1 x1 + + al,n xn = bl
and
(cai,1 )x1 + + (cai,n )xn

= c(ai,1 x1 + + ai,n xn )
= cbi

Thus, (x1 , . . . , xn ) is a solution of

8
>
>
>
>
>
>
<
>
>
>
>
>
>
:

a1,1 x1
..
.

am,1 x1

c(ai,1 x1 ) +
..
.

a1,n xn
..
.

a1,n xn
..
.

b1
..
.

+ c(ai,n xn ) = cbi
..
..
.
.
+ am,n xn = bm

Conversely, let (x1 , . . . , xn ) be a solution of

8
>
>
>
>
>
>
<
>
>
>
>
>
>
:

a1,1 x1
..
.

am,1 x1

c(ai,1 x1 ) +
..
.

b1
..
.

+ c(ai,n xn ) = cbi
..
..
.
.
+ am,n xn = bm

For all l 2 {1, . . . , m} so that l 6= k,

al,1 x1 + + al,n xn = bl
and
ai,1 x1 + + ai,n xn

= (1/c)(c(ai,1 x1 ) + + c(ai,n xn ))
= (1/c)cbi
= bi

79

3.1. SYSTEMS OF LINEAR EQUATIONS

Thus, (x1 , . . . , xn ) is a solution of
8
a1,1 x1
>
>
>
>
..
>
>
.
<
ai,1 x1
>
>
..
>
>
>
.
>
:
am,1 x1

+ am,n xn

a1,n xn
..
.

ai,n xn
..
.

b1
..
.
bi
..
.

= bm

Therefore, the two linear systems are equivalent.

(ii) Exercise.
The following examples illustrate this approach to solving linear systems.
Example 3.1.9. Solve the system
8
2y
< x
x + 7y
:
2x
y

7z
+ 17z
6z

=
=
=

16
36
14

-2

0
2

3
z

2
1

2
0
y
2

Figure 3.1: Example 3.1.9

Solution. Adding the left- and right-hand sides of equation 1 to those of equation 2,
8
2y
7z =
16
< x
5y + 10z =
20
:
2x
y
6z =
14
Subtracting twice the left- and right-hand sides of
8
2y
< x
5y +
:
3y +

equation 1 from those of equation 3,

7z
10z
8z

=
=
=

16
20
18