SECTION 3.

F Duality 101

3.F Duality
The Dual Space and the Dual Map
Linear maps into the scalar field F play a special role in linear algebra, and
thus they get a special name:

3.92 Definition linear functional

A linear functional on V is a linear map from V to F. In other words, a
linear functional is an element of L.V; F/.

3.93 Example linear functionals

 Define ' W R3 ! R by '.x; y; z/ D 4x  5y C 2z. Then ' is a linear

functional on R3 .

 Fix .c1 ; : : : ; cn / 2 Fn . Define ' W Fn ! F by

'.x1 ; : : : ; xn / D c1 x1 C    C cn xn :
Then ' is a linear functional on Fn .

 Define ' W P.R/ ! R by '.p/ D 3p 00 .5/ C 7p.4/. Then ' is a linear

functional on P.R/.
R1
 Define ' W P.R/ ! R by '.p/ D 0 p.x/ dx. Then ' is a linear
functional on P.R/.

The vector space L.V; F/ also gets a special name and special notation:

3.94 Definition dual space, V 0

The dual space of V, denoted V 0, is the vector space of all linear
functionals on V. In other words, V 0 D L.V; F/.

3.95 dim V 0 D dim V

Suppose V is finite-dimensional. Then V 0 is also finite-dimensional and
dim V 0 D dim V.

Proof This result follows from 3.61.

102 CHAPTER 3 Linear Maps

In the following definition, 3.5 implies that each 'j is well defined.

3.96 Definition dual basis

If v1 ; : : : ; vn is a basis of V, then the dual basis of v1 ; : : : ; vn is the list
'1 ; : : : ; 'n of elements of V 0, where each 'j is the linear functional on V
such that (
1 if k D j;
'j .vk / D
0 if k ¤ j:

3.97 Example What is the dual basis of the standard basis e1 ; : : : ; en

of Fn ?

Solution For 1  j  n, define 'j to be the linear functional on Fn that

selects the j th coordinate of a vector in Fn . In other words,
'j .x1 ; : : : ; xn / D xj
for .x1 ; : : : ; xn / 2 Fn . Clearly
(
1 if k D j;
'j .ek / D
0 if k ¤ j:
Thus '1 ; : : : ; 'n is the dual basis of the standard basis e1 ; : : : ; en of Fn .

The next result shows that the dual basis is indeed a basis. Thus the
terminology “dual basis” is justified.

3.98 Dual basis is a basis of the dual space

Suppose V is finite-dimensional. Then the dual basis of a basis of V is a
basis of V 0.

Proof Suppose v1 ; : : : ; vn is a basis of V. Let '1 ; : : : ; 'n denote the dual

basis.
To show that '1 ; : : : ; 'n is a linearly independent list of elements of V 0,
suppose a1 ; : : : ; an 2 F are such that
a1 '1 C    C an 'n D 0:
Now .a1 '1 C    C an 'n /.vj / D aj for j D 1; : : : ; n. The equation
above thus shows that a1 D    D an D 0. Hence '1 ; : : : ; 'n is linearly
independent.
Now 2.39 and 3.95 imply that '1 ; : : : ; 'n is a basis of V 0.
 SECTION 3.F Duality 103

In the definition below, note that if T is a linear map from V to W then T 0

is a linear map from W 0 to V 0.

3.99 Definition dual map, T 0

If T 2 L.V; W /, then the dual map of T is the linear map T 0 2 L.W 0 ; V 0 /
defined by T 0 .'/ D ' ı T for ' 2 W 0.

If T 2 L.V; W / and ' 2 W 0, then T 0 .'/ is defined above to be the

composition of the linear maps ' and T. Thus T 0 .'/ is indeed a linear map
from V to F; in other words, T 0 .'/ 2 V 0.
The verification that T 0 is a linear map from W 0 to V 0 is easy:

 If '; 2 W 0, then

T 0 .' C / D .' C /ıT D'ıT C ı T D T 0 .'/ C T 0 . /:

 If  2 F and ' 2 W 0, then

T 0 .'/ D .'/ ı T D .' ı T / D T 0 .'/:

In the next example, the prime notation is used with two unrelated mean-
ings: D 0 denotes the dual of a linear map D, and p 0 denotes the derivative of
a polynomial p.

3.100 Example Define D W P.R/ ! P.R/ by Dp D p 0.

 Suppose ' is the linear functional on P.R/ defined by '.p/ D p.3/.

Then D 0 .'/ is the linear functional on P.R/ given by
 0 
D .'/ .p/ D .' ı D/.p/ D '.Dp/ D '.p 0 / D p 0 .3/:
In other words, D 0 .'/ is the linear functional on P.R/ that takes p to
p 0 .3/.
R1
 Suppose ' is the linear functional on P.R/ defined by '.p/ D 0 p.
Then D 0 .'/ is the linear functional on P.R/ given by
Z 1
 0  0
D .'/ .p/ D .'ıD/.p/ D '.Dp/ D '.p / D p 0 D p.1/p.0/:
0
In other words, D 0 .'/ is the linear functional on P.R/ that takes p to
p.1/  p.0/.
104 CHAPTER 3 Linear Maps

The first two bullet points in the result below imply that the function that
takes T to T 0 is a linear map from L.V; W / to L.W 0 ; V 0 /.
In the third bullet point below, note the reversal of order from ST on the
left to T 0 S 0 on the right (here we assume that U is a vector space over F).

3.101 Algebraic properties of dual maps

 .S C T /0 D S 0 C T 0 for all S; T 2 L.V; W /.

 .T /0 D T 0 for all  2 F and all T 2 L.V; W /.

 .S T /0 D T 0 S 0 for all T 2 L.U; V / and all S 2 L.V; W /.

Proof The proofs of the first two bullet points above are left to the reader.
To prove the third bullet point, suppose ' 2 W 0. Then
 
.ST /0 .'/ D ' ı.S T / D .' ıS /ıT D T 0 .' ıS / D T 0 S 0 .'/ D .T 0 S 0 /.'/;

Some books use the notation V  where the first, third, and fourth equal-
and T  for duality instead of V 0 ities above hold because of the defini-
and T 0. However, here we reserve tion of the dual map, the second equality
the notation T  for the adjoint, holds because composition of functions
which will be introduced when we is associative, and the last equality fol-
study linear maps on inner product lows from the definition of composition.
spaces in Chapter 7.
The equality of the first and last
terms above for all ' 2 W 0 means that
.S T /0 D T 0 S 0.

The Null Space and Range of the Dual of a Linear Map

Our goal in this subsection is to describe null T 0 and range T 0 in terms of
range T and null T. To do this, we will need the following definition.

3.102 Definition annihilator, U 0

For U  V, the annihilator of U, denoted U 0 , is defined by
U 0 D f' 2 V 0 W '.u/ D 0 for all u 2 U g:

3.103 Example Suppose U is the subspace of P.R/ consisting of all

polynomial multiples of x 2 . If ' is the linear functional on P.R/ defined by
'.p/ D p 0 .0/, then ' 2 U 0 .
 SECTION 3.F Duality 105

For U  V, the annihilator U 0 is a subset of the dual space V 0. Thus U 0

depends on the vector space containing U, so a notation such as UV0 would be
more precise. However, the containing vector space will always be clear from
the context, so we will use the simpler notation U 0 .

3.104 Example Let e1 ; e2 ; e3 ; e4 ; e5 denote the standard basis of R5 , and

0
let '1 ; '2 ; '3 ; '4 ; '5 denote the dual basis of .R5 / . Suppose

U D span.e1 ; e2 / D f.x1 ; x2 ; 0; 0; 0/ 2 R5 W x1 ; x2 2 Rg:

Show that U 0 D span.'3 ; '4 ; '5 /.

Solution Recall (see 3.97) that 'j is the linear functional on R5 that selects
that j th coordinate: 'j .x1 ; x2 ; x3 ; x4 ; x5 / D xj .
First suppose ' 2 span.'3 ; '4 ; '5 /. Then there exist c3 ; c4 ; c5 2 R such
that ' D c3 '3 C c4 '4 C c5 '5 . If .x1 ; x2 ; 0; 0; 0/ 2 U, then

'.x1 ; x2 ; 0; 0; 0/ D .c3 '3 C c4 '4 C c5 '5 /.x1 ; x2 ; 0; 0; 0/ D 0:

Thus ' 2 U 0 . In other words, we have shown that span.'3 ; '4 ; '5 /  U 0 .
To show the inclusion in the other direction, suppose ' 2 U 0 . Because
0
the dual basis is a basis of .R5 / , there exist c1 ; c2 ; c3 ; c4 ; c5 2 R such that
' D c1 '1 C c2 '2 C c3 '3 C c4 '4 C c5 '5 . Because e1 2 U and ' 2 U 0 , we
have

0 D '.e1 / D .c1 '1 C c2 '2 C c3 '3 C c4 '4 C c5 '5 /.e1 / D c1 :

Similarly, e2 2 U and thus c2 D 0. Hence ' D c3 '3 C c4 '4 C c5 '5 . Thus

' 2 span.'3 ; '4 ; '5 /, which shows that U 0  span.'3 ; '4 ; '5 /.

3.105 The annihilator is a subspace

Suppose U  V. Then U 0 is a subspace of V 0.

Proof Clearly 0 2 U 0 (here 0 is the zero linear functional on V ), because

the zero linear functional applied to every vector in U is 0.
Suppose '; 2 U 0 . Thus '; 2 V 0 and '.u/ D .u/ D 0 for every
u 2 U. If u 2 U, then .' C /.u/ D '.u/ C .u/ D 0 C 0 D 0. Thus
' C 2 U 0.
Similarly, U 0 is closed under scalar multiplication. Thus 1.34 implies that
U 0 is a subspace of V 0.
106 CHAPTER 3 Linear Maps

The next result shows that dim U 0 is the difference of dim V and dim U.
For example, this shows that if U is a 2-dimensional subspace of R5 , then U 0
0
is a 3-dimensional subspace of .R5 / , as in Example 3.104.
The next result can be proved following the pattern of Example 3.104:
choose a basis u1 ; : : : ; um of U, extend to a basis u1 ; : : : ; um ; : : : ; un of V,
let '1 ; : : : ; 'm ; : : : ; 'n be the dual basis of V 0, and then show 'mC1 ; : : : ; 'n
is a basis of U 0 , which implies the desired result.
You should construct the proof outlined in the paragraph above, even
though a slicker proof is presented here.

3.106 Dimension of the annihilator

Suppose V is finite-dimensional and U is a subspace of V. Then

dim U C dim U 0 D dim V:

Proof Let i 2 L.U; V / be the inclusion map defined by i.u/ D u for u 2 U.

Thus is a linear map from V 0 to U 0. The Fundamental Theorem of Linear
i0
Maps (3.22) applied to i 0 shows that

dim range i 0 C dim null i 0 D dim V 0 :

However, null i 0 D U 0 (as can be seen by thinking about the definitions) and
dim V 0 D dim V (by 3.95), so we can rewrite the equation above as

dim range i 0 C dim U 0 D dim V:

If ' 2 U 0, then ' can be extended to a linear functional on V (see,

for example, Exercise 11 in Section 3.A). The definition of i 0 shows that
i 0 . / D '. Thus ' 2 range i 0, which implies that range i 0 D U 0. Hence
dim range i 0 D dim U 0 D dim U, and the displayed equation above becomes
the desired result.

The proof of part (a) of the result below does not use the hypothesis that
V and W are finite-dimensional.

3.107 The null space of T 0

Suppose V and W are finite-dimensional and T 2 L.V; W /. Then

(a) null T 0 D .range T /0 ;

(b) dim null T 0 D dim null T C dim W  dim V.

 SECTION 3.F Duality 107

Proof

(a) First suppose ' 2 null T 0. Thus 0 D T 0 .'/ D ' ı T. Hence

0 D .' ı T /.v/ D '.T v/ for every v 2 V:

Thus ' 2 .range T /0 . This implies that null T 0  .range T /0 .

To prove the inclusion in the opposite direction, now suppose that
' 2 .range T /0 . Thus '.T v/ D 0 for every vector v 2 V. Hence
0 D ' ı T D T 0 .'/. In other words, ' 2 null T 0, which shows that
.range T /0  null T 0, completing the proof of (a).

(b) We have

dim null T 0 D dim.range T /0

D dim W  dim range T
D dim W  .dim V  dim null T /
D dim null T C dim W  dim V;

where the first equality comes from (a), the second equality comes from
3.106, and the third equality comes from the Fundamental Theorem of
Linear Maps (3.22).

The next result can be useful because sometimes it is easier to verify that
T0 is injective than to show directly that T is surjective.

3.108 T surjective is equivalent to T 0 injective

Suppose V and W are finite-dimensional and T 2 L.V; W /. Then T is
surjective if and only if T 0 is injective.

Proof The map T 2 L.V; W / is surjective if and only if range T D W,

which happens if and only if .range T /0 D f0g, which happens if and only if
null T 0 D f0g [by 3.107(a)], which happens if and only if T 0 is injective.

3.109 The range of T 0

Suppose V and W are finite-dimensional and T 2 L.V; W /. Then

(a) dim range T 0 D dim range T ;

(b) range T 0 D .null T /0 .

108 CHAPTER 3 Linear Maps

Proof

(a) We have

dim range T 0 D dim W 0  dim null T 0

D dim W  dim.range T /0
D dim range T;

where the first equality comes from the Fundamental Theorem of Linear
Maps (3.22), the second equality comes from 3.95 and 3.107(a), and
the third equality comes from 3.106.

(b) First suppose ' 2 range T 0. Thus there exists 2 W 0 such that
' D T 0 . /. If v 2 null T, then
 
'.v/ D T 0 . / v D . ı T /.v/ D .T v/ D .0/ D 0:

Hence ' 2 .null T /0 . This implies that range T 0  .null T /0 .

We will complete the proof by showing that range T 0 and .null T /0
have the same dimension. To do this, note that

dim range T 0 D dim range T

D dim V  dim null T
D dim.null T /0 ;

where the first equality comes from (a), the second equality comes from
the Fundamental Theorem of Linear Maps (3.22), and the third equality
comes from 3.106.

The next result should be compared to 3.108.

3.110 T injective is equivalent to T 0 surjective

Suppose V and W are finite-dimensional and T 2 L.V; W /. Then T is
injective if and only if T 0 is surjective.

Proof The map T 2 L.V; W / is injective if and only if null T D f0g,

which happens if and only if .null T /0 D V 0, which happens if and only if
range T 0 D V 0 [by 3.109(b)], which happens if and only if T 0 is surjective.
 SECTION 3.F Duality 109

The Matrix of the Dual of a Linear Map

We now define the transpose of a matrix.

3.111 Definition transpose, At

The transpose of a matrix A, denoted At , is the matrix obtained from
A by interchanging the rows and columns. More specifically, if A is an
m-by-n matrix, then At is the n-by-m matrix whose entries are given by
the equation
.At /k;j D Aj;k :

0 1
5 7  
@ A 5 3 4
3.112 Example If A D 3 8 , then A D
t .
7 8 2
4 2
Note that here A is a 3-by-2 matrix and At is a 2-by-3 matrix.

The transpose has nice algebraic properties: .A C C /t D At C C t and

.A/t D At for all m-by-n matrices A; C and all  2 F (see Exercise 33).
The next result shows that the transpose of the product of two matrices is
the product of the transposes in the opposite order.

3.113 The transpose of the product of matrices

If A is an m-by-n matrix and C is an n-by-p matrix, then

.AC /t D C t At :

Proof Suppose 1  k  p and 1  j  m. Then

 
.AC /t k;j D .AC /j;k
X
n
D Aj;r Cr;k
rD1
Xn
D .C t /k;r .At /r;j
rD1
D .C t At /k;j :

Thus .AC /t D C t At , as desired.

110 CHAPTER 3 Linear Maps

The setting for the next result is the assumption that we have a basis
v1 ; : : : ; vn of V, along with its dual basis '1 ; : : : ; 'n of V 0. We also have a
basis w1 ; : : : ; wm of W, along with its dual basis 1 ; : : : ; m of W 0. Thus
M.T / is computed with respect to the bases just mentioned of V and W,
and M.T 0 / is computed with respect to the dual bases just mentioned of W 0
and V 0.

3.114 The matrix of T 0 is the transpose of the matrix of T

 t
Suppose T 2 L.V; W /. Then M.T 0 / D M.T / .

Proof Let A D M.T / and C D M.T 0 /. Suppose 1  j  m and

1  k  n.
From the definition of M.T 0 / we have

X
n
T 0. j/ D Cr;j 'r :
rD1

The left side of the equation above equals j ı T. Thus applying both sides
of the equation above to vk gives

X
n
. j ı T /.vk / D Cr;j 'r .vk /
rD1
D Ck;j :

We also have

. j ı T /.vk / D j .T vk /
Xm 
D j Ar;k wr
rD1
X
m
D Ar;k j .wr /
rD1
D Aj;k :

Comparing the last line of the last two sets of equations, we have Ck;j D Aj;k .
 t
Thus C D At . In other words, M.T 0 / D M.T / , as desired.
 SECTION 3.F Duality 111

The Rank of a Matrix

We begin by defining two nonnegative integers that are associated with each
matrix.

3.115 Definition row rank, column rank

Suppose A is an m-by-n matrix with entries in F.

 The row rank of A is the dimension of the span of the rows of A in

F1;n .

 The column rank of A is the dimension of the span of the columns

of A in Fm;1 .

 
4 7 1 8
3.116 Example Suppose A D . Find the row rank of A
3 5 2 9
and the column rank of A.

Solution The row rank of A is the dimension of

   
span 4 7 1 8 ; 3 5 2 9

in F1;4 . Neither of the two vectors listed above in F1;4 is a scalar multiple
of the other. Thus the span of this list of length 2 has dimension 2. In other
words, the row rank of A is 2.
The column rank of A is the dimension of
       !
4 7 1 8
span ; ; ;
3 5 2 9

in F2;1 . Neither of the first two vectors listed above in F2;1 is a scalar multiple
of the other. Thus the span of this list of length 4 has dimension at least 2.
The span of this list of vectors in F2;1 cannot have dimension larger than 2
because dim F2;1 D 2. Thus the span of this list has dimension 2. In other
words, the column rank of A is 2.

Notice that no bases are in sight in the statement of the next result. Al-
though M.T / in the next result depends on a choice of bases of V and W,
the next result shows that the column rank of M.T / is the same for all such
choices (because range T does not depend on a choice of basis).
112 CHAPTER 3 Linear Maps

3.117 Dimension of range T equals column rank of M.T /

Suppose V and W are finite-dimensional and T 2 L.V; W /. Then
dim range T equals the column rank of M.T /.

Proof Suppose v1 ; : : : ; vn is a basis of V and w1 ; : : : ; wm is a basis of W. The

function that takes w 2 span.T v1 ; : : : ; T vn / to M.w/  is easily seen to be an
isomorphism from span.T v1 ; : : : ; T vn / onto  span M.T v1 /; : : : ; M.T  vn / .
Thus dim span.T v1 ; : : : ; T vn / D dim span M.T v1 /; : : : ; M.T vn / , where
the last dimension equals the column rank of M.T /.
It is easy to see that range T D span.T v1 ; : : : ; T vn /. Thus we have
dim range T D dim span.T v1 ; : : : ; T vn / D the column rank of M.T /, as
desired.

In Example 3.116, the row rank and column rank turned out to equal each
other. The next result shows that this always happens.

3.118 Row rank equals column rank

Suppose A 2 Fm;n . Then the row rank of A equals the column rank of A.

Proof Define T W Fn;1 ! Fm;1 by T x D Ax. Thus M.T / D A, where

M.T / is computed with respect to the standard bases of Fn;1 and Fm;1 . Now
column rank of A D column rank of M.T /
D dim range T
D dim range T 0
D column rank of M.T 0 /
D column rank of At
D row rank of A;
where the second equality above comes from 3.117, the third equality comes
from 3.109(a), the fourth equality comes from 3.117 (where M.T 0 / is com-
puted with respect to the dual bases of the standard bases), the fifth equality
comes from 3.114, and the last equality follows easily from the definitions.

The last result allows us to dispense with the terms “row rank” and “column
rank” and just use the simpler term “rank”.

3.119 Definition rank

The rank of a matrix A 2 Fm;n is the column rank of A.
 SECTION 3.F Duality 113

EXERCISES 3.F

1 Explain why every linear functional is either surjective or the zero map.

2 Give three distinct examples of linear functionals on RŒ0;1 .

3 Suppose V is finite-dimensional and v 2 V with v ¤ 0. Prove that there

exists ' 2 V 0 such that '.v/ D 1.

4 Suppose V is finite-dimensional and U is a subspace of V such that

U ¤ V. Prove that there exists ' 2 V 0 such that '.u/ D 0 for every
u 2 U but ' ¤ 0.

5 Suppose V1 ; : : : ; Vm are vector spaces. Prove that .V1      Vm /0 and

V1 0      Vm 0 are isomorphic vector spaces.

6 Suppose V is finite-dimensional and v1 ; : : : ; vm 2 V. Define a linear

map  W V 0 ! Fm by
 
.'/ D '.v1 /; : : : ; '.vm / :

(a) Prove that v1 ; : : : ; vm spans V if and only if  is injective.

(b) Prove that v1 ; : : : ; vm is linearly independent if and only if  is
surjective.

7 Suppose m is a positive integer. Show that the dual basis of the basis
.j /
1; x; : : : ; x m of Pm .R/ is '0 ; '1 ; : : : ; 'm , where 'j .p/ D p j Š.0/ . Here
p .j / denotes the j th derivative of p, with the understanding that the 0th
derivative of p is p.

8 Suppose m is a positive integer.

(a) Show that 1; x  5; : : : ; .x  5/m is a basis of Pm .R/.

(b) What is the dual basis of the basis in part (a)?

9 Suppose v1 ; : : : ; vn is a basis of V and '1 ; : : : ; 'n is the corresponding

dual basis of V 0. Suppose 2 V 0. Prove that

D .v1 /'1 C    C .vn /'n :

10 Prove the first two bullet points in 3.101.

114 CHAPTER 3 Linear Maps

11 Suppose A is an m-by-n matrix with A ¤ 0. Prove that the rank of A

is 1 if and only if there exist .c1 ; : : : ; cm / 2 Fm and .d1 ; : : : ; dn / 2 Fn
such that Aj;k D cj dk for every j D 1; : : : ; m and every k D 1; : : : ; n.

12 Show that the dual map of the identity map on V is the identity map
on V 0.

13 Define T W R3 ! R2 by T .x; y; z/ D .4x C 5y C 6z; 7x C 8y C 9z/.

Suppose '1 ; '2 denotes the dual basis of the standard basis of R2 and
3
1 ; 2 ; 3 denotes the dual basis of the standard basis of R .

(a) Describe the linear functionals T 0 .'1 / and T 0 .'2 /.

(b) Write T 0 .'1 / and T 0 .'2 / as linear combinations of 1; 2; 3.

14 Define T W P.R/ ! P.R/ by .Tp/.x/ D x 2 p.x/ C p 00 .x/ for x 2 R.

(a) Suppose ' 2 P.R/0 is defined by '.p/ D p 0 .4/. Describe the

linear functional T 0 .'/ on P.R/.
0
R1
(b)  0  '32 P.R/ is defined by '.p/ D 0 p.x/ dx. Evaluate
Suppose
T .'/ .x /.

15 Suppose W is finite-dimensional and T 2 L.V; W /. Prove that T 0 D 0

if and only if T D 0.

16 Suppose V and W are finite-dimensional. Prove that the map that takes
T 2 L.V; W / to T 0 2 L.W 0 ; V 0 / is an isomorphism of L.V; W / onto
L.W 0 ; V 0 /.

17 Suppose U  V. Explain why U 0 D f' 2 V 0 W U  null 'g.

18 Suppose V is finite-dimensional and U  V. Show that U D f0g if and

only if U 0 D V 0.

19 Suppose V is finite-dimensional and U is a subspace of V. Show that

U D V if and only if U 0 D f0g.

20 Suppose U and W are subsets of V with U  W. Prove that W 0  U 0 .

21 Suppose V is finite-dimensional and U and W are subspaces of V with

W 0  U 0 . Prove that U  W.

22 Suppose U; W are subspaces of V. Show that .U C W /0 D U 0 \ W 0 .

 SECTION 3.F Duality 115

23 Suppose V is finite-dimensional and U and W are subspaces of V. Prove

that .U \ W /0 D U 0 C W 0 .

24 Prove 3.106 using the ideas sketched in the discussion before the state-
ment of 3.106.

25 Suppose V is finite-dimensional and U is a subspace of V. Show that

U D fv 2 V W '.v/ D 0 for every ' 2 U 0 g:

26 Suppose V is finite-dimensional and  is a subspace of V 0. Show that

 D fv 2 V W '.v/ D 0 for every ' 2 g0 :

 
27 Suppose T 2 L P5 .R/; P5 .R/ and null T 0 D span.'/, where ' is
the linear functional on P5 .R/ defined by '.p/ D p.8/. Prove that
range T D fp 2 P5 .R/ W p.8/ D 0g.

28 Suppose V and W are finite-dimensional, T 2 L.V; W /, and there exists

' 2 W 0 such that null T 0 D span.'/. Prove that range T D null '.

29 Suppose V and W are finite-dimensional, T 2 L.V; W /, and there exists

' 2 V 0 such that range T 0 D span.'/. Prove that null T D null '.

30 Suppose V is finite-dimensional and '1 ; : : : ; 'm is a linearly independent

list in V 0. Prove that
 
dim .null '1 / \    \ .null 'm / D .dim V /  m:

31 Suppose V is finite-dimensional and '1 ; : : : ; 'n is a basis of V 0. Show

that there exists a basis of V whose dual basis is '1 ; : : : ; 'n .

32 Suppose T 2 L.V /, and u1 ; : : : ; un and v1 ; : : : ; vn are bases of V. Prove

that the following are equivalent:

(a) T is invertible.
(b) The columns of M.T / are linearly independent in Fn;1 .
(c) The columns of M.T / span Fn;1 .
(d) The rows of M.T / are linearly independent in F1;n .
(e) The rows of M.T / span F1;n .
 
Here M.T / means M T; .u1 ; : : : ; un /; .v1 ; : : : ; un / .
116 CHAPTER 3 Linear Maps

33 Suppose m and n are positive integers. Prove that the function that takes
A to At is a linear map from Fm;n to Fn;m . Furthermore, prove that this
linear map is invertible.
34 The double dual space of V, denoted V 00, is defined to be the dual space
of V 0. In other words, V 00 D .V 0 /0. Define ƒ W V ! V 00 by
.ƒv/.'/ D '.v/
for v 2 V and ' 2 V 0.
(a) Show that ƒ is a linear map from V to V 00.
(b) Show that if T 2 L.V /, then T 00 ı ƒ D ƒ ı T, where T 00 D .T 0 /0.
(c) Show that if V is finite-dimensional, then ƒ is an isomorphism
from V onto V 00.
[Suppose V is finite-dimensional. Then V and V 0 are isomorphic, but
finding an isomorphism from V onto V 0 generally requires choosing a
basis of V. In contrast, the isomorphism ƒ from V onto V 00 does not
require a choice of basis and thus is considered more natural.]
 0
35 Show that P.R/ and R1 are isomorphic.
36 Suppose U is a subspace of V. Let i W U ! V be the inclusion map
defined by i.u/ D u. Thus i 0 2 L.V 0 ; U 0 /.
(a) Show that null i 0 D U 0 .
(b) Prove that if V is finite-dimensional, then range i 0 D U 0.
(c) Prove that if V is finite-dimensional, then ei 0 is an isomorphism
0 0
from V =U onto U . 0

[The isomorphism in part (c) is natural in that it does not depend on a

choice of basis in either vector space.]
37 Suppose U is a subspace
 of V. Let  W V ! V =U be the usual quotient
map. Thus  0 2 L .V =U /0 ; V 0 .
(a) Show that  0 is injective.
(b) Show that range  0 D U 0 .
(c) Conclude that  0 is an isomorphism from .V =U /0 onto U 0 .
[The isomorphism in part (c) is natural in that it does not depend on a
choice of basis in either vector space. In fact, there is no assumption
here that any of these vector spaces are finite-dimensional.]