In this section we shall begin the study of functions of the form ,

4.2 where the independent variable x is a vector in and the dependent variable
LINEAR w is a vector in . We shall concentrate on a special class of such functions
TRANSFORMATIONS called “linear transformations.” Linear transformations are fundamental in the
study of linear algebra and have many important applications in physics,
FROM Rn TO Rm engineering, social sciences, and various branches of mathematics.

Functions from to R

Recall that a function is a rule f that associates with each element in a set A one and only one element in a set B. If f associates
the element b with the element a, then we write and say that b is the image of a under f or that is the value of f at
a. The set A is called the domain of f and the set B is called the codomain of f. The subset of B consisting of all possible values
for f as a varies over A is called the range of f. For the most common functions, A and B are sets of real numbers, in which case f
is called a real-valued function of a real variable. Other common functions occur when B is a set of real numbers and A is a set
of vectors in , , or, more generally, . Some examples are shown in Table 1. Two functions and are regarded as
equal, written , if they have the same domain and for all a in the domain.

Table 1

Formula Example Classification Description

Real-valued function of a real variable Function from R to R

Real-valued function of two real Function from to

variables R

Real-valued function of three real Function from to

variables R

Real-valued function of n real variables Function from to

R

Functions from to

If the domain of a function f is and the codomain is (m and n possibly the same), then f is called a map or transformation
from to , and we say that the function f maps into . We denote this by writing . The functions in
Table 1 are transformations for which . In the case where , the transformation is called an operator on
. The first entry in Table 1 is an operator on R.

To illustrate one important way in which transformations can arise, suppose that , , …, are real-valued functions of n
real variables, say

(1)

These m equations assign a unique point in to each point in and thus define a
transformation from to . If we denote this transformation by T, then and

EXAMPLE 1 A Transformation from to

The equations

define a transformation . With this transformation, the image of the point is

Thus, for example,

Linear Transformations from to

In the special case where the equations in 1 are linear, the transformation defined by those equations is called a
linear transformation (or a linear operator if ). Thus a linear transformation is defined by equations of the
form

(2)

or, in matrix notation,

(3)

or more briefly by

(4)

The matrix is called the standard matrix for the linear transformation T, and T is called multiplication by A.

EXAMPLE 2 A Linear Transformation from to

The linear transformation defined by the equations

(5)

can be expressed in matrix form as

 (6)

so the standard matrix for T is

The image of a point can be computed directly from the defining equations 5 or from 6 by matrix multiplication.
For example, if , then substituting in 5 yields

(verify) or alternatively from 6,

Some Notational Matters

If is multiplication by A, and if it is important to emphasize that A is the standard matrix for T, we shall denote the
linear transformation by . Thus

(7)

It is understood in this equation that the vector x in is expressed as a column matrix.

Sometimes it is awkward to introduce a new letter to denote the standard matrix for a linear transformation . In
such cases we will denote the standard matrix for T by the symbol .With this notation, equation 7 would take the form

(8)

Occasionally, the two notations for a standard matrix will be mixed, in which case we have the relationship

(9)

Remark Amidst all of this notation, it is important to keep in mind that we have established a correspondence between
matrices and linear transformations from to : To each matrix A there corresponds a linear transformation
(multiplication by A), and to each linear transformation , there corresponds an matrix (the standard
matrix for T).

Geometry of Linear Transformations

Depending on whether n-tuples are regarded as points or vectors, the geometric effect of an operator is to
transform each point (or vector) in into some new point (or vector) (Figure 4.2.1).
 Figure 4.2.1

EXAMPLE 3 Zero Transformation from to

If 0 is the zero matrix and 0 is the zero vector in , then for every vector x in ,

so multiplication by zero maps every vector in into the zero vector in . We call the zero transformation from to
. Sometimes the zero transformation is denoted by 0. Although this is the same notation used for the zero matrix, the appropriate
interpretation will usually be clear from the context.

EXAMPLE 4 Identity Operator on

If I is the identity matrix, then for every vector x in ,

so multiplication by I maps every vector in into itself. We call the identity operator on . Sometimes the identity
operator is denoted by I. Although this is the same notation used for the identity matrix, the appropriate interpretation will
usually be clear from the context.

Among the most important linear operators on and are those that produce reflections, projections, and rotations. We shall
now discuss such operators.

Reflection Operators
Consider the operator that maps each vector into its symmetric image about the y-axis (Figure 4.2.2).

Figure 4.2.2

If we let , then the equations relating the components of x and w are

(10)

or, in matrix form,

(11)

Since the equations in 10 are linear, T is a linear operator, and from 11 the standard matrix for T is

In general, operators on and that map each vector into its symmetric image about some line or plane are called reflection
operators. Such operators are linear. Tables 2 and 3 list some of the common reflection operators.

Table 2

Operator Illustration Equations Standard Matrix

Reflection about the y-axis

Reflection about the x-axis

Reflection about the line

Table 3
 Operator Illustration Equations Standard Matrix

Reflection about the -plane

Reflection about the -plane

Reflection about the -plane

Projection Operators

Consider the operator that maps each vector into its orthogonal projection on the x-axis (Figure 4.2.3). The
equations relating the components of x and are

(12)

or, in matrix form,

(13)

Figure 4.2.3

The equations in 12 are linear, so T is a linear operator, and from 13 the standard matrix for T is

In general, a projection operator (more precisely, an orthogonal projection operator) on or is any operator that maps
each vector into its orthogonal projection on a line or plane through the origin. It can be shown that such operators are linear.
Some of the basic projection operators on and are listed in Tables 4 and 5.
 Table 4

Operator Illustration Equations Standard Matrix

Orthogonal projection on the x-axis

Orthogonal projection on the y-axis

Table 5

Operator Illustration Equations Standard Matrix

Orthogonal projection on the -plane

Orthogonal projection on the -plane

Orthogonal projection on the -plane

Rotation Operators

An operator that rotates each vector in through a fixed angle is called a rotation operator on . Table 6 gives the formula
for the rotation operators on . To show how this is derived, consider the rotation operator that rotates each vector
counterclockwise through a fixed positive angle . To find equations relating x and , let be the angle from the
positive x-axis to x, and let r be the common length of x and w (Figure 4.2.4).
 Figure 4.2.4

Table 6

Operator Illustration Equations Standard Matrix

Rotation through an angle

Then, from basic trigonometry,

(14)

and

(15)

Using trigonometric identities on 15 yields

and substituting14 yields

(16)

The equations in 16 are linear, so T is a linear operator; moreover, it follows from these equations that the standard matrix for T
is

EXAMPLE 5 Rotation

If each vector in is rotated through an angle of , then the image w of a vector

is
For example, the image of the vector

A rotation of vectors in is usually described in relation to a ray emanating from the origin, called the axis of rotation. As a
vector revolves around the axis of rotation, it sweeps out some portion of a cone (Figure 4.2.5a). The angle of rotation, which is
measured in the base of the cone, is described as “clockwise” or “counterclockwise” in relation to a viewpoint that is along the
axis of rotation looking toward the origin. For example, in Figure 4.2.5a the vector w results from rotating the vector x
counterclockwise around the axis l through an angle . As in , angles are positive if they are generated by counterclockwise
rotations and negative if they are generated by clockwise rotations.

Figure 4.2.5

The most common way of describing a general axis of rotation is to specify a nonzero vector u that runs along the axis of
rotation and has its initial point at the origin. The counterclockwise direction for a rotation about the axis can then be
determined by a “right-hand rule” (Figure 4.2.5b): If the thumb of the right hand points in the direction of u, then the cupped
fingers point in a counterclockwise direction.

A rotation operator on is a linear operator that rotates each vector in about some rotation axis through a fixed angle . In
Table 7 we have described the rotation operators on whose axes of rotation are the positive coordinate axes. For each of
these rotations one of the components is unchanged by the rotation, and the relationships between the other components can be
derived by the same procedure used to derive 16. For example, in the rotation about the z-axis, the z-components of x and
are the same, and the x- and y-components are related as in 16. This yields the rotation equation shown in the last row
of Table 7.

Table 7
Operator Illustration Equations Standard Matrix

Counterclockwise rotation about the

positive x-axis through an angle

Counterclockwise rotation about the

positive y-axis through an angle

Counterclockwise rotation about the

positive z-axis through an angle

Yaw, Pitch, and Roll

In aeronautics and astronautics, the orientation of an aircraft or space shuttle relative to an -coordinate system is often
described in terms of angles called yaw, pitch, and roll. If, for example, an aircraft is flying along the y-axis and the -plane
defines the horizontal, then the aircraft's angle of rotation about the z-axis is called the yaw, its angle of rotation about the
x-axis is called the pitch, and its angle of rotation about the y-axis is called the roll. A combination of yaw, pitch, and roll can
be achieved by a single rotation about some axis through the origin. This is, in fact, how a space shuttle makes attitude
adjustments—it doesn't perform each rotation separately; it calculates one axis, and rotates about that axis to get the correct
orientation. Such rotation maneuvers are used to align an antenna, point the nose toward a celestial object, or position a
payload bay for docking.
For completeness, we note that the standard matrix for a counterclockwise rotation through an angle about an axis in ,
which is determined by an arbitrary unit vector that has its initial point at the origin, is

(17)

The derivation can be found in the book Principles of Interactive Computer Graphics, by W. M. Newman and R. F. Sproull
(New York: McGraw-Hill, 1979). The reader may find it instructive to derive the results in Table 7 as special cases of this more
general result.

Dilation and Contraction Operators

If k is a nonnegative scalar, then the operator on or is called a contraction with factor k if and a
dilation with factor k if . The geometric effect of a contraction is to compress each vector by a factor of k (Figure 4.2.6a),
and the effect of a dilation is to stretch each vector by a factor of k (Figure 4.2.6b). A contraction compresses or uniformly
toward the origin from all directions, and a dilation stretches or uniformly away from the origin in all directions.

Figure 4.2.6

The most extreme contraction occurs when , in which case reduces to the zero operator , which
compresses every vector into a single point (the origin). If , then reduces to the identity operator ,
which leaves each vector unchanged; this can be regarded as either a contraction or a dilation. Tables 8 and 9 list the dilation
and contraction operators on and .

Table 8

Operator Illustration Equations Standard Matrix

 Operator Illustration Equations Standard Matrix

Contraction with factor k on

Dilation with factor k on

Table 9

Operator Illustration Equations Standard Matrix

Contraction with factor k on

Dilation with factor k on

Rotations in

A familiar example of a rotation in is the rotation of the Earth about its axis through the North and South Poles. For
simplicity, we will assume that the Earth is a sphere. Since the Sun rises in the east and sets in the west, we know that the
Earth rotates from west to east. However, to an observer above the North Pole the rotation will appear counterclockwise, and
 to an observer below the South Pole it will appear clockwise. Thus, when a rotation in is described as clockwise or
counterclockwise, a direction of view along the axis of rotation must also be stated.

There are some other facts about the Earth's rotation that are useful for understanding general rotations in . For example, as
the Earth rotates about its axis, the North and South Poles remain fixed, as do all other points that lie on the axis of rotation.
Thus, the axis of rotation can be thought of as the line of fixed points in the Earth's rotation. Moreover, all points on the
Earth that are not on the axis of rotation move in circular paths that are centered on the axis and lie in planes that are
perpendicular to the axis. For example, the points in the Equatorial Plane move within the Equatorial Plane in circles about
the Earth's center.

Compositions of Linear Transformations

If and are linear transformations, then for each x in one can first compute , which is a
vector in , and then one can compute , which is a vector in . Thus, the application of followed by
produces a transformation from to . This transformation is called the composition of with and is denoted by
(read “ circle ”). Thus

(18)

The composition is linear since

(19)

so is multiplication by , which is a linear transformation. Formula 19 also tells us that the standard matrix for
is . This is expressed by the formula

(20)

Remark Formula 20 captures an important idea: Multiplying matrices is equivalent to composing the corresponding linear
transformations in the right-to-left order of the factors.

There is an alternative form of Formula 20: If and are linear transformations, then because the
standard matrix for the composition is the product of the standard matrices of and T, we have

(21)

EXAMPLE 6 Composition of Two Rotations

Let and be the linear operators that rotate vectors through the angles and , respectively. Thus
the operation

first rotates x through the angle , then rotates through the angle . It follows that the net effect of is to rotate
each vector in through the angle (Figure 4.2.7).
 Figure 4.2.7

Thus the standard matrices for these linear operators are

These matrices should satisfy 21. With the help of some basic trigonometric identities, we can show that this is so as follows:

Remark In general, the order in which linear transformations are composed matters. This is to be expected, since the
composition of two linear transformations corresponds to the multiplication of their standard matrices, and we know that the
order in which matrices are multiplied makes a difference.

EXAMPLE 7 Composition Is Not Commutative

Let be the reflection operator about the line , and let be the orthogonal projection on the
y-axis. Figure 4.2.8 illustrates graphically that and have different effects on a vector x. This same conclusion
can be reached by showing that the standard matrices for and do not commute:

so .
 Figure 4.2.8

EXAMPLE 8 Composition of Two Reflections

Let be the reflection about the y-axis, and let be the reflection about the x-axis. In this case
and are the same; both map each vector into its negative (Figure 4.2.9):

Figure 4.2.9

The equality of and can also be deduced by showing that the standard matrices for and commute:
The operator on or is called the reflection about the origin. As the computations above show, the standard
matrix for this operator on is

Compositions of Three or More Linear Transformations

Compositions can be defined for three or more linear transformations. For example, consider the linear transformations

We define the composition by

It can be shown that this composition is a linear transformation and that the standard matrix for is related to the
standard matrices for , , and by

(22)

which is a generalization of 21. If the standard matrices for , , and are denoted by A, B, and C, respectively, then we also
have the following generalization of 20:

(23)

EXAMPLE 9 Composition of Three Transformations

Find the standard matrix for the linear operator that first rotates a vector counterclockwise about the z-axis
through an angle , then reflects the resulting vector about the -plane, and then projects that vector orthogonally onto the
-plane.

Solution

The linear transformation T can be expressed as the composition

where is the rotation about the z-axis, is the reflection about the -plane, and is the orthogonal projection on the
-plane. From Tables 3, 5, and 7, the standard matrices for these linear transformations are

Thus, from 22 the standard matrix for T is ; that is,

 Exercise Set 4.2

Click here for Just Ask!

Find the domain and codomain of the transformation defined by the equations, and determine whether the transformation is
1. linear.

(a)

(b)

(c)

(d)

Find the standard matrix for the linear transformation defined by the equations.
2.

(a)

(b)
 (c)

(d)

Find the standard matrix for the linear operator given by

3.

and then calculate by directly substituting in the equations and also by matrix multiplication.

Find the standard matrix for the linear operator T defined by the formula.
4.

(a)

(b)

(c)

(d)

Find the standard matrix for the linear transformation T defined by the formula.
5.

(a)

(b)

(c)

(d)

In each part, the standard matrix of a linear transformation T is given. Use it to find . [Express the answers in
6. matrix form.]
 (a)

(b)

(c)

(d)

In each part, use the standard matrix for T to find ; then check the result by calculating directly.
7.

(a) ;

(b) ;

Use matrix multiplication to find the reflection of (−1, 2) about

8.

(a) the x-axis

(b) the y-axis

(c) the line

Use matrix multiplication to find the reflection of (2, −5, 3) about

9.

(a) the -plane

(b) the -plane

(c) the -plane

 Use matrix multiplication to find the orthogonal projection of (2, −5) on
10.

(a) the x-axis

(b) the y-axis

Use matrix multiplication to find the orthogonal projection of (−2, 1, 3) on

11.

(a) the -plane

(b) the -plane

(c) the -plane

Use matrix multiplication to find the image of the vector (3, −4) when it is rotated through an angle of
12.

(a)

(b)

(c)

(d)

Use matrix multiplication to find the image of the vector (−2, 1, 2) if it is rotated
13.

(a) 30° about the x-axis

(b) 45° about the y-axis

(c) 90° about the z-axis

Find the standard matrix for the linear operator that rotates a vector in through an angle of about
14.

(a) the x-axis

(b) the y-axis

 (c) the z-axis

Use matrix multiplication to find the image of the vector (−2, 1, 2) if it is rotated
15.

(a) about the x-axis

(b) about the y-axis

(c) about the z-axis

Find the standard matrix for the stated composition of linear operators on .
16.

(a) A rotation of 90°, followed by a reflection about the line .

(b) An orthogonal projection on the y-axis, followed by a contraction with factor .

(c) A reflection about the x-axis, followed by a dilation with factor .

Find the standard matrix for the stated composition of linear operators on .
17.

(a) A rotation of 60°, followed by an orthogonal projection on the x-axis, followed by a reflection about the line .

(b) A dilation with factor , followed by a rotation of 45°, followed by a reflection about the y-axis.

(c) A rotation of 15°, followed by a rotation of 105°, followed by a rotation of 60°.

Find the standard matrix for the stated composition of linear operators on .
18.

(a) A reflection about the -plane, followed by an orthogonal projection on the -plane.

(b) A rotation of 45° about the y-axis, followed by a dilation with factor .

(c) An orthogonal projection on the -plane, followed by a reflection about the -plane.
 Find the standard matrix for the stated composition of linear operators on .
19.

(a) A rotation of 30° about the x-axis, followed by a rotation of 30° about the z-axis, followed by a contraction with
factor .

(b) A reflection about the -plane, followed by a reflection about the -plane, followed by an orthogonal projection
on the -plane.

(c) A rotation of 270° about the x-axis, followed by a rotation of 90° about the y-axis, followed by a rotation of 180°
about the z-axis.

Determine whether .
20.

(a) is the orthogonal projection on the x-axis, and is the orthogonal projection on the
y-axis.

(b) is the rotation through an angle , and is the rotation through an angle .

(c) is the orthogonal projection on the x-axis, and is the rotation through an angle .

Determine whether .
21.

(a) is a dilation by a factor k, and is the rotation about the z-axis

through an angle .

(b) is the rotation about the x-axis through an angle , and is the rotation about the z-axis
through an angle .

In the orthogonal projections on the x-axis, y-axis, and z-axis are defined by
22.

respectively.

(a) Show that the orthogonal projections on the coordinate axes are linear operators, and find their standard matrices.

(b) Show that if is an orthogonal projection on one of the coordinate axes, then for every vector x in
the vectors and are orthogonal vectors.
 (c) Make a sketch showing x and in the case where T is the orthogonal projection on the x-axis.

Derive the standard matrices for the rotations about the x-axis, y-axis, and z-axis in from Formula 17.
23.

Use Formula 17 to find the standard matrix for a rotation of radians about the axis determined by the vector
24. .

Note Formula 17 requires that the vector defining the axis of rotation have length 1.

Verify Formula 21 for the given linear transformations.

25.

(a) and

(b) and

(c) and

It can be proved that if A is a matrix with and such that the column vectors of A are orthogonal and have
26. length 1, then multiplication by A is a rotation through some angle . Verify that

satisfies the stated conditions and find the angle of rotation.

The result stated in Exercise 26 is also true in : It can be proved that if A is a matrix with and such that
27. the column vectors of A are pairwise orthogonal and have length 1, then multiplication by A is a rotation about some axis of
rotation through some angle . Use Formula 17 to show that if A satisfies the stated conditions, then the angle of rotation
satisfies the equation

Let A be a matrix (other than the identity matrix) satisfying the conditions stated in Exercise 27. It can be shown tha
28. if x is any nonzero vector in , then the vector determines an axis of rotation when u is
positioned with its initial point at the origin. [See “The Axis of Rotation: Analysis, Algebra, Geometry,” by Dan Kalman,
Mathematics Magazine, Vol. 62, No. 4, October 1989.]

(a) Show that multiplication by

 is a rotation.

(b) Find a vector of length 1 that defines an axis for the rotation.

(c) Use the result in Exercise 27 to find the angle of rotation about the axis obtained in part (b).

In words, describe the geometric effect of multiplying a vector x by the matrix A.

29.

(a)

(b)

In words, describe the geometric effect of multiplying a vector x by the matrix A.

30.

(a)

(b)

In words, describe the geometric effect of multiplying a vector x by the matrix

31.

If multiplication by A rotates a vector x in the -plane through an angle , what is the effect of
32. multiplying x by ? Explain your reasoning.

Let be a nonzero column vector in , and suppose that is the transformation

33. defined by , where is the standard matrix of the rotation of about the origin
through the angle . Give a geometric description of this transformation. Is it a linear
transformation? Explain.

A function of the form is commonly called a “linear function” because the graph
34. of is a line. Is f a linear transformation on R?

Let be a line in , and let be a linear operator on . What kind of

35. geometric object is the image of this line under the operator T? Explain your reasoning.
Copyright © 2005 John Wiley & Sons, Inc. All rights reserved.
 4.3 In this section we shall investigate the relationship between the invertibility of a
matrix and properties of the corresponding matrix transformation. We shall also
PROPERTIES OF LINEAR obtain a characterization of linear transformations from to that will form
TRANSFORMATIONS the basis for more general linear transformations to be discussed in subsequent
FROM TO sections, and we shall discuss some geometric properties of eigenvectors.

One-to-One Linear Transformations

Linear transformations that map distinct vectors (or points) into distinct vectors (or points) are of special importance. One example
of such a transformation is the linear operator that rotates each vector through an angle . It is obvious
geometrically-that if u and v are distinct vectors in , then so are the rotated vectors and (Figure 4.3.1).

Figure 4.3.1
Distinct vectors u and v are rotated into distinct vectors and .

In contrast, if is the orthogonal projection of on the -plane, then distinct points on the same vertical line are
mapped into the same point in the -plane (Figure 4.3.2).

Figure 4.3.2
The distinct points P and Q are mapped into the same point M.

DEFINITION

A linear transformation is said to be one-to-one if T maps distinct vectors (points) in into distinct vectors
(points) in .

Remark It follows from this definition that for each vector w in the range of a one-to-one linear transformation T, there is exactly
one vector x such that .
EXAMPLE 1 One-to-One Linear Transformations

In the terminology of the preceding definition, the rotation operator of Figure 4.3.1 is one-to-one, but the orthogonal projection
operator of Figure 4.3.2 is not.

Let A be an matrix, and let be multiplication by A. We shall now investigate relationships between the
invertibility of A and properties of .

Recall from Theorem 2.3.6 (with w in place of b) that the following are equivalent:

A is invertible.

is consistent for every matrix w.

has exactly one solution for every matrix w.

However, the last of these statements is actually stronger than necessary. One can show that the following are equivalent (Exercise
24):

A is invertible.

is consistent for every matrix w.

has exactly one solution when the system is consistent.

Translating these into the corresponding statements about the linear operator , we deduce that the following are equivalent:

A is invertible.

For every vector w in , there is some vector x in such that . Stated another way, the range of is all of .

For every vector w in the range of , there is exactly one vector x in such that . Stated another way, is
one-to-one.

In summary, we have established the following theorem about linear operators on .

THEOREM 4.3.1

Equivalent Statements

If A is an matrix and is multiplication by A, then the following statements are equivalent.

 (a) A is invertible.

(b) The range of is .

(c) is one-to-one.

EXAMPLE 2 Applying Theorem 4.3.1

In Example 1 we observed that the rotation operator illustrated in Figure 4.3.1 is one-to-one. It follows from
Theorem 4.3.1 that the range of T must be all of and that the standard matrix for T must be invertible. To show that the range of
T is all of , we must show that every vector w in is the image of some vector x under T. But this is clearly so, since the vector
x obtained by rotating w through the angle maps into w when rotated through the angle . Moreover, from Table 6 of Section
4.2, the standard matrix for T is

which is invertible, since

EXAMPLE 3 Applying Theorem 4.3.1

In Example 1 we observed that the projection operator illustrated in Figure 4.3.2 is not one-to-one. It follows from
Theorem 4.3.1 that the range of T is not all of and that the standard matrix for T is not invertible. To show directly that the
range of T is not all of , we must find a vector w in that is not the image of any vector x under T. But any vector w outside of
the -plane has this property, since all images under T lie in the -plane. Moreover, from Table 5 of Section 4.2, the standard
matrix for T is

which is not invertible, since .

Inverse of a One-to-One Linear Operator

If is a one-to-one linear operator, then from Theorem 4.3.1 the matrix A is invertible. Thus, is
itself a linear operator; it is called the inverse of . The linear operators and cancel the effect of one another in the sense
that for all x in ,

or, equivalently,
From a more geometric viewpoint, if w is the image of x under , then maps w back into x, since

(Figure 4.3.3).

Figure 4.3.3

Before turning to an example, it will be helpful to touch on a notational matter. When a one-to-one linear operator on is written
as (rather than ), then the inverse of the operator T is denoted by (rather than ). Since the
standard matrix for is the inverse of the standard matrix for T, we have

(1)

EXAMPLE 4 Standard Matrix for

Let be the operator that rotates each vector in through the angle , so from Table 6 of Section 4.2,

(2)

It is evident geometrically that to undo the effect of T, one must rotate each vector in through the angle . But this is exactly
what the operator does, since the standard matrix for is

(verify), which is identical to 2 except that is replaced by .

EXAMPLE 5 Finding

Show that the linear operator defined by the equations

is one-to-one, and find .

Solution

The matrix form of these equations is

so the standard matrix for T is

This matrix is invertible (so T is one-to-one) and the standard matrix for is

Thus

from which we conclude that

Linearity Properties

In the preceding section we defined a transformation to be linear if the equations relating x and are linear
equations. The following theorem provides an alternative characterization of linearity. This theorem is fundamental and will be the
basis for extending the concept of a linear transformation to more general settings later in this text.

THEOREM 4.3.2

Properties of Linear Transformations

A transformation is linear if and only if the following relationships hold for all vectors u and v in and for
every scalar c.

(a)

(b)

Proof Assume first that T is a linear transformation, and let A be the standard matrix for T. It follows from the basic arithmetic
properties of matrices that

and

Conversely, assume that properties (a) and (b) hold for the transformation T. We can prove that T is
linear by finding a matrix A with the property that
 (3)

for all vectors x in . This will show that T is multiplication by A and therefore linear. But before we
can produce this matrix, we need to observe that property (a) can be extended to three or more terms;
for example, if u, v, and w are any vectors in , then by first grouping v and w and applying property
(a), we obtain

More generally, for any vectors , , …, in , we have

Now, to find the matrix A, let , , …, be the vectors

(4)

and let A be the matrix whose successive column vectors are , , …, ; that is,

(5)

If

is any vector in , then as discussed in Section 1.3, the product is a linear combination of the
column vectors of A with coefficients from x, so

which completes the proof.

Expression 5 is important in its own right, since it provides an explicit formula for the standard matrix of a linear operator
in terms of the images of the vectors , , …, under T. For reasons that will be discussed later, the vectors ,
, …, in 4 are called the standard basis vectors for . In and these are the vectors of length 1 along the coordinate axes
(Figure 4.3.4).
 Figure 4.3.4

Because of its importance, we shall state 5 as a theorem for future reference.

THEOREM 4.3.3

If is a linear transformation, and , , …, are the standard basis vectors for , then the standard matrix
for T is

(6)

Formula 6 is a powerful tool for finding standard matrices and analyzing the geometric effect of a linear transformation. For
example, suppose that is the orthogonal projection on the -plane. Referring to Figure 4.3.4, it is evident
geometrically that

so by 6,

which agrees with the result in Table 5 of Section 4.2.

Using 6 another way, suppose that is multiplication by

The images of the standard basis vectors can be read directly from the columns of the matrix A:
EXAMPLE 6 Standard Matrix for a Projection Operator

Let l be the line in the -plane that passes through the origin and makes an angle with the positive x-axis, where . As
illustrated in Figure 4.3.5a, let be a linear operator that maps each vector into its orthogonal projection on l.

Figure 4.3.5

(a) Find the standard matrix for T.

(b) Find the orthogonal projection of the vector onto the line through the origin that makes an angle of with
the positive x-axis.

Solution (a)
From 6,

where and are the standard basis vectors for . We consider the case where ; the case where is
similar. Referring to Figure 4.3.5b, we have , so

and referring to Figure 4.3.5c, we have , so

Thus the standard matrix for T is

Solution (b)

Since and , it follows from part (a) that the standard matrix for this projection operator is

Thus

or, in point notation,

Geometric Interpretation of Eigenvectors

Recall from Section 2.3 that if A is an matrix, then is called an eigenvalue of A if there is a nonzero vector x such that

The nonzero vectors x satisfying this equation are called the eigenvectors of A corresponding to .

Eigenvalues and eigenvectors can also be defined for linear operators on ; the definitions parallel those for matrices.

DEFINITION

If is a linear operator, then a scalar is called an eigenvalue of T if there is a nonzero x in such that

(7)

Those nonzero vectors x that satisfy this equation are called the eigenvectors of T corresponding to .
Observe that if A is the standard matrix for T, then 7 can be written as

from which it follows that

The eigenvalues of T are precisely the eigenvalues of its standard matrix A.

x is an eigenvector of T corresponding to if and only if x is an eigenvector of A corresponding to .

If is an eigenvalue of A and x is a corresponding eigenvector, then , so multiplication by A maps x into a scalar multiple
of itself. In and , this means that multiplication by A maps each eigenvector x into a vector that lies on the same line as x
(Figure 4.3.6).

Figure 4.3.6

Recall from Section 4.2 that if , then the linear operator compresses x by a factor of if or stretches x by a
factor of if . If , then reverses the direction of x and compresses the reversed vector by a factor of if
or stretches the reversed vector by a factor of if (Figure 4.3.7).

Figure 4.3.7

EXAMPLE 7 Eigenvalues of a Linear Operator

Let be the linear operator that rotates each vector through an angle . It is evident geometrically that unless is a
multiple of , T does not map any nonzero vector x onto the same line as x; consequently, T has no real eigenvalues. But if is a
multiple of , then every nonzero vector x is mapped onto the same line as x, so every nonzero vector is an eigenvector of T. Let us
verify these geometric observations algebraically. The standard matrix for T is

As discussed in Section 2.3, the eigenvalues of this matrix are the solutions of the characteristic equation

that is,

(8)

But if is not a multiple of , then , so this equation has no real solution for , and consequently A has no real
eigenvalues.* If is a multiple of , then and either or , depending on the particular multiple of . In
the case where and , the characteristic equation 8 becomes , so is the only eigenvalue of A. In
this case the matrix A is

Thus, for all x in ,

so T maps every vector to itself, and hence to the same line. In the case where and , the characteristic equation
8 becomes , so is the only eigenvalue of A. In this case the matrix A is

Thus, for all x in ,

so T maps every vector to its negative, and hence to the same line as x.

EXAMPLE 8 Eigenvalues of a Linear Operator

Let be the orthogonal projection on the -plane. Vectors in the -plane are mapped into themselves under T, so
each nonzero vector in the -plane is an eigenvector corresponding to the eigenvalue . Every vector x along the z-axis is
mapped into 0 under T, which is on the same line as x, so every nonzero vector on the z-axis is an eigenvector corresponding to the
eigenvalue . Vectors that are not in the -plane or along the z-axis are not mapped into scalar multiples of themselves, so
there are no other eigenvectors or eigenvalues.

To verify these geometric observations algebraically, recall from Table 5 of Section 4.2 that the standard matrix for T is

The characteristic equation of A is

which has the solutions and anticipated above.

As discussed in Section 2.3, the eigenvectors of the matrix A corresponding to an eigenvalue are the nonzero solutions of
 (9)

If , this system is

which has the solutions , , (verify), or, in matrix form,

As anticipated, these are the vectors along the z-axis. If , then system 9 is

which has the solutions , , (verify), or, in matrix form,

As anticipated, these are the vectors in the -plane.

Summary

In Theorem 2.3.6 we listed six results that are equivalent to the invertibility of a matrix A. We conclude this section by merging
Theorem 4.3.1 with that list to produce the following theorem that relates all of the major topics we have studied thus far.

THEOREM 4.3.4

Equivalent Statements

If A is an matrix, and if is multiplication by A, then the following are equivalent.

(a) A is invertible.

(b) has only the trivial solution.

(c) The reduced row-echelon form of A is .

(d) A is expressible as a product of elementary matrices.

(e) is consistent for every matrix b.

(f) has exactly one solution for every matrix b.

 (g) .

(h) The range of is .

(i) is one-to-one.

Exercise Set 4.3

Click here for Just Ask!

By inspection, determine whether the linear operator is one-to-one.

1.

(a) the orthogonal projection on the x-axis in

(b) the reflection about the y-axis in

(c) the reflection about the line in

(d) a contraction with factor in

(e) a rotation about the z-axis in

(f) a reflection about the -plane in

(g) a dilation with factor in

Find the standard matrix for the linear operator defined by the equations, and use Theorem 4.3.4 to determine whether the
2. operator is one-to-one.

(a)

(b)
 (c)

(d)

Show that the range of the linear operator defined by the equations
3.

is not all of , and find a vector that is not in the range.

Show that the range of the linear operator defined by the equations
4.

is not all of , and find a vector that is not in the range.

Determine whether the linear operator defined by the equations is one-to-one; if so, find the standard matrix for
5. the inverse operator, and find .

(a)

(b)

(c)

(d)

Determine whether the linear operator defined by the equations is one-to-one; if so, find the standard matrix for
6. the inverse operator, and find .

(a)
 (b)

(c)

(d)

By inspection, determine the inverse of the given one-to-one linear operator.

7.

(a) the reflection about the x-axis in

(b) the rotation through an angle of in

(c) the dilation by a factor of 3 in

(d) the reflection about the -plane in

(e) the contraction by a factor of in

In Exercises 8 and 9 use Theorem 4.3.2 to determine whether is a linear operator.

8.
(a)

(b)

(c)

(d)

9.
(a)
 (b)

(c)

(d)

In Exercises 10 and 11 use Theorem 4.3.2 to determine whether is a linear transformation.

10.
(a)

(b)

11.
(a)

(b)

In each part, use Theorem 4.3.3 to find the standard matrix for the linear operator from the images of the standard basis
12. vectors.

(a) the reflection operators on in Table 2 of Section 4.2

(b) the reflection operators on in Table 3 of Section 4.2

(c) the projection operators on in Table 4 of Section 4.2

(d) the projection operators on in Table 5 of Section 4.2

(e) the rotation operators on in Table 6 of Section 4.2

(f) the dilation and contraction operators on in Table 9 of Section 4.2

Use Theorem 4.3.3 to find the standard matrix for from the images of the standard basis vectors.
13.

(a) projects a vector orthogonally onto the x-axis and then reflects that vector about the y-axis.
 (b) reflects a vector about the line and then reflects that vector about the x-axis.

(c) dilates a vector by a factor of 3, then reflects that vector about the line , and then projects that
vector orthogonally onto the y-axis.

Use Theorem 4.3.3 to find the standard matrix for from the images of the standard basis vectors.
14.

(a) reflects a vector about the -plane and then contracts that vector by a factor of .

(b) projects a vector orthogonally onto the -plane and then projects that vector orthogonally onto the
-plane.

(c) reflects a vector about the -plane, then reflects that vector about the -plane, and then reflects that
vector about the -plane.

Let be multiplication by
15.

and let , , and be the standard basis vectors for . Find the following vectors by inspection.

(a) , , and

(b)

(c)

Determine whether multiplication by A is a one-to-one linear transformation.

16.

(a)

(b)
 (c)

Use the result in Example 6 to find the orthogonal projection of x onto the line through the origin that makes an angle with
17. the positive x-axis.

(a) ;

(b) ;

(c) ;

Use the type of argument given in Example 8 to find the eigenvalues and corresponding eigenvectors of T. Check your
18. conclusions by calculating the eigenvalues and corresponding eigenvectors from the standard matrix for T.

(a) is the reflection about the x-axis.

(b) is the reflection about the line .

(c) is the orthogonal projection on the x-axis.

(d) is the contraction by a factor of .

Follow the directions of Exercise 18.

19.

(a) is the reflection about the -plane.

(b) is the orthogonal projection on the -plane.

(c) is the dilation by a factor of 2.

(d) is a rotation of about the z-axis.

20.
(a) Is a composition of one-to-one linear transformations one-to-one? Justify your conclusion.

(b) Can the composition of a one-to-one linear transformation and a linear transformation that is not one-to-one be
one-to-one? Account for both possible orders of composition and justify your conclusion.

Show that defines a linear operator on but does not.

21.

22.
(a) Prove that if is a linear transformation, then —that is, T maps the zero vector in into the
zero vector in .

(b) The converse of this is not true. Find an example of a function that satisfies but is not a linear transformation.

Let l be the line in the -plane that passes through the origin and makes an angle with the positive x-axis, where .
23. Let be the linear operator that reflects each vector about l (see the accompanying figure).

(a) Use the method of Example 6 to find the standard matrix for T.

(b) Find the reflection of the vector about the line l through the origin that makes an angle of with the
positive x-axis.

Figure Ex-23

Prove: An matrix A is invertible if and only if the linear system has exactly one solution for every vector w in
24. for which the system is consistent.

Indicate whether each statement is always true or sometimes false. Justify your answer by giving a
25. logical argument or a counterexample.

(a) If T maps into , and , then T is linear.

 (b) If is a one-to-one linear transformation, then there are no distinct vectors u
and v in such that .

(c) If is a linear operator, and if for some vector x, then is an

eigenvalue of T.

(d) If T maps into , and if for all scalars and and

for all vectors u and v in , then T is linear.

Indicate whether each statement is always true, sometimes true, or always false.
26.

(a) If is a linear transformation and , then T is one-to-one.

(b) If is a linear transformation and , then T is one-to-one.

(c) If is a linear transformation and , then T is one-to-one.

Let A be an matrix such that , and let be multiplication by A.

27.

(a) What can you say about the range of the linear operator T? Give an example that illustrates
your conclusion.

(b) What can you say about the number of vectors that T maps into 0?

In each part, make a conjecture about the eigenvectors and eigenvalues of the matrix A
28. corresponding to the given transformation by considering the geometric properties of multiplication
by A. Confirm each of your conjectures with computations.

(a) Reflection about the line .

(b) Contraction by a factor of .

Copyright © 2005 John Wiley & Sons, Inc. All rights reserved.