CHAPTER 5

GENERAL VECTOR
SPACES
 CONTENTS
5.1. Real Vector Spaces
5.2. Subspaces
5.3. Linear Independence
5.4. Basis and Dimension
5.5. Row Space, Column Space, and Null
Space
5.6. Rank and Nullity
5.1. REAL VECTOR SPACES

5.1.1. Vector Space Axioms

5.1.2. Examples of Vector Spaces
5.1.3. Properties of Vector Spaces
5.1.1. VECTOR SPACE AXIOMS

Let V be a nonempty set. Assume that there exist two

operations on V:
- The addition operation that associates to each pair of
elements u, v in V an element of V denoted by u + v,
called the sum of u and v.
- The scalar multiplication that associates to each
element u in V and a scalar k an element of V denoted
by ku, called the scalar multiple of u by k
- The elements of V are called vectors
 Definition. A vector space is a set V together with two
operations: addition and scalar multiplication, such
that the following axioms are satisfied for all u, v ,
w  V and all scalars k, l
(1) u + v = v + u
(2) (u + v) + w = u + (v + w)
(3) There exists a vector 0 such that u + 0 = u
(4) For all uV, there is –uV such that u + (–u) = 0
(5) k(l u) = (kl)u
(6) k(u + v) = ku + kv
(7) (k + l)u = k u + l u
(8) 1u = u
If the set of scalars is R, V is a real vector space
If the set of scalars is C, V is a complex vector space
5.1.2. EXAMPLES OF VECTOR SPACES

Example. Rn is a real vector space. In particular R2 and

R3 are real vector spaces

Example. The set Mmn of m  n matrices with the

operations of matrix addition, scalar multiplication is
a real vector space
The vector 0 is the zero matrix
Example. The zero vector space consists of a single
vector 0 with the addition: 0 + 0 = 0
and scalar multiplication: k0 = 0
Example. Let F(–,) be the set of all real-valued
functions defined on R with the pointwise addition,
scalar multiplication operations:
(f + g)(x) = f(x) + g(x)
(kf)(x) = kf(x)
Then F(–,) is a real vector space
The vector 0 is the zero function, and the negative
vector –f is the function
(–f)(x) = –f(x)

Example. Let V be a plane through the origin in R3

with the equation:
ax + by + cz = 0
 ax + by + cz = 0
Let u = (u1, u2, u3) and v = (v1, v2, v3) be two
vectors in V, we have
au1 + bu2 + cu3 = 0
av1 + bv2 + cv3 = 0
It follows that
a(u1+ v1) + b(u2 + v2) + c(u3 + v3)= 0
Thus u + v is also a vector in V
Similarly ku, 0 and the negative vector –u are in V.
The other axioms are obviously true. This shows that
V is a real vector space called a subspace of R3
5.1.3. PROPERTIES OF VECTORS SPACE

Theorem. Let V be a vector space, u a vector in V, and

k a scalar, then
(a) 0u = 0
(b) k0 = 0
(c) (–1)u = –u
(d) If ku = 0, then k = 0 or u = 0
 5.2. SUBSPACES

5.2.1. Subspaces
5.2.2. Solution Spaces of Homogeneous
Systems
5.2.3. Spanning
 5.2.1. SUBSPACES

Definition. A subspace of a vector space V is a subset

W of V such that W is itself a vector space under the
addition and scalar multiplication defined on V

Theorem. If W is a subset of a vector space V, then W

is a subspace of V if and only if the following
conditions hold
(a) If u and v are vectors in W then u + v is in W
(b) If k is any scalar and u is any vector in W then ku
is in W
If the conditions (a) and (b) hold, we say that W is
closed under the addition and scalar multiplication

Example. Let V be a plane through the origin in R3, we

have proved that V is closed under addition and scalar
multiplication. Therefore by the Theorem,
it is not necessary to verify the other axioms to conclude
that V is a subspace of R3.
In particular V is itself a vector space
Similarly we can verify easily that any line through the
origin is closed under addition and scalar multiplication,
and hence is a subspace of R3
Example. Let W be the first quadrant of R2:
W = {(x,y) / x  0 and y  0 }

Then we can verify that W is closed under addition but

not under scalar multiplication.
For example v = (1, 1) lies in W but (–1)v = (–1, –1)
does not

Example. Let Sn be the set of n  n symmetric

matrices then Sn is closed under matrix addition and
scalar multiplication.
Therefore Sn is subspace of the vector space Mnn of all
n  n matrices
Example. Let Pn be the set of all polynomials of degree  n
in one variable.
Then Pn, as a subset of the vector space F(–, ) of all
real-valued functions defined on R, is clearly closed
under pointwise addition and scalar multiplication:
(p + q)(x) = p(x) + q(x)
(kp)(x) = kp(x)
Therefore Pn is a subspace of F(–, )

Example. Let C(–, ) be the set of all continuous

functions on R. Then C(–, ) is also closed under
pointwise addition and scalar multiplication.
Hence it is a subspace of F(–, )
Example. Similarly let Cm(–, ) be the set of all
functions on R having continuous derivatives up to
order m.
Then Cm(–, ) is also a subspace of C(–, ).
The same is true for C(–, ), the subset of those
functions having continuous derivatives of all order.
We have
C(–, )  …  Cm+1 (–, ) 
 Cm(–, )  …  C(–, )
If we consider only the functions defined on a
closed or open finite interval I. Then we have the
sequence of vector spaces
C(I)  …  Cm+1 (I)  Cm(I)  …  C(I)
 5.2.2. Solution Spaces of
Homogeneous Systems
Theorem. If Au = 0 is a homogeneous linear system of
m equations in n unknowns, then the set of solutions
is a subspace W of Rn.

Proof. W contains clearly 0. On the other hand let x

and x' be two solutions and k be a scalar, then
A(x + x') = Ax + Ax' = 0
A(kx) = kAx = 0

Therefore W is a subspace of Rn
Example. Consider the linear system
 1 − 2 3  x  0 
 2 − 4 6   y  = 0 
    
3 − 6 9  z  0

Then the solutions are

x = 2s – 3t, y = s, z=t
It follows that
x = 2y – 3z or x –2y + 3z = 0
Therefore the solution set is the plane through the
origin with n = (1, –2, 3) as a normal
Example. Consider the linear system
 1 − 2 3   x  0 
 − 3 7 − 8   y  = 0 
    
− 2 4 − 6  z  0

Then the solutions are

x = –5t, y = –t, z=t

Therefore the solution set is the line through the

origin parallel to the vector v = (–5, –1, 1)
Example. Consider the linear system
 1 − 2 3   x  0 
− 3 7 − 8  y  = 0
    
 4 1 2   z  0
Then the solution is x = 0, y = 0, z = 0
Therefore the solution set is the origin only, i.e. {0}
Example. Consider the linear system
 0 0 0   x  0 
 0 0 0   y  = 0 
    
0 0 0  z  0
Then every vector is a solution, and hence the
solution set is R3
5.2.3. SPANNING
From the proof of the last Theorem, we also see that if
x1, x2, …, xr are solutions of the equation Ax = 0,
then the following linear combination is also a solution

k1x1 + k2x2 + … + krxr

In fact this is true for any subspace W of a vector

space V: if v1, v2, …, vr are elements in W,
then the following linear combination is also in W
k1v1 + k2v2 + … + krvr
Note. The vector v = (b1, b2, …, bn) is a linear
combination of the vectors v1= (a11, a21, …, an1),
v2 = (a12, a22, …, an2), …, vr = (a1r, a2r, …, anr)
if and only if there are scalar k1, k2, …, kn such that

k1v1+ k2 v2+ … + kr vr = v

i.e. (k1, k2, …, kn) is a solution of the linear system

k1a11+ k2a12+ … + kra1r = b1

k1a21+ k2a22+ … + kra2r = b2
… … …
k1an1+ k2an2+ … + kranr = bn
Example. The vector w = (9, 2, 7) is a linear combination
of the vectors u = (1, 2, –1), and v = (6, 4, 2) since the
linear system
k1+ 6k2 = 9
2k1+ 4k2 = 2
–k1+ 2k2 = 7

has a solution: k1= –3, k2 = 2

We have
w = –3u + 2v
Example. The vector w = (4, –1, 8 ) is not a linear
combination of the vectors u = (1, 2, –1), and v = (6, 4, 2)
since the linear system

k1+ 6k2 = 4
2k1+ 4k2 = –1
–k1+ 2k2 = 8

is inconsistent
Adding indeed the last equation and the negative of
the 1st equation to the 2nd equation we get
0k1+ 0k2 = 3
Theorem. If v1, v2, …, vr are elements of a vector
space V, then:
(a) The set W of all linear combinations of v1, v2, …,
vr is a subspace of V
(b) W is the smallest subspace of V that contains v1,
v2, …, vr in the sense that every subspace of V that
contains v1, v2, …, vr must contain W

The set W of all linear combinations of v1, v2, …, vr is

called the subspace of V spanned by v1, v2, …, vr and
denoted by
span{v1, v2, …, vr}
Example. Let v1, and v2 are two non-colinear vectors
in R3 with their initial points at the origin.

Then span{v1, v2} consists of all linear combinations

k1v1+ k2v2.
Therefore it is precisely the plane determined by v1
and v2
k1v1+ k2v2

v2

v1
Similarly let v be a nonzero vector in R2 or R3.

Then span{v} consists of all scalar multiple kv

Therefore it is precisely the line kv

determined by v
v
Example. The polynomials
1, x, x2, …, xn span the vector
space Pn
since each polynomial p in Pn can be written as a
linear combination

p = a0 + a1x + a2x2+ … + anxn

Example. Determine whether the vectors v1=(1,1,2),
v2=(1,0,1), and v3=(2,1,3) span R3

Solution. Note that v1, v2, and v3 span R3 if and only

if the following linear system has a solution for all
b=(b1, b2, b3) in R3
k1 + k2 + 2k3 = b1
k1 + k3 = b2
2k1+ k2 + 3k3 = b3

i.e. the coefficient matrix is invertible.

 However
1 1 2
1 0 1 =0
2 1 3

Therefore v1, v2, and v3 do not span R3

Note. The spanning sets of a vector space are not unique.

For instance a line through the origin in R2 or R3 is
spanned by any nonzero vectors lying on this line
5.3. LINEAR INDEPENDENCE

5.3.1. Linearly Independent sets

5.3.2. Geometric Interpretation of Linear
Independence
5.3.3. Linear Independence of Functions
 5.3.1. Linearly Independent sets

Definition. A nonempty set of vectors S = {v1, v2, …, vr}

in a vector space V is said to be a linearly independent
set if the following vector equation
k1v1 + k2v2 + … + krvr = 0
has a unique solution k1= 0, k2= 0, …, kr= 0.
Otherwise S is said to be a linearly dependent set
Example. The vectors v1=(2,–1,0,3), v2=(1,2,5,–1)
and v3 = (7,–1,5,8) are linearly dependent since

3v1 + v2 –v3 = 0

Example. The polynomials

p1 = 1 – x, p2 = 5 + 3x – 2x2 and p3 = 1 + 3x – x2

are linearly dependent since

3p1–p2 + 2p3 = 0
 Example. Let i, j and k are the standard basis vectors
of R3. Consider the equation
k1i + k2j + k3k = 0
or
k1(1,0,0) + k2 (0,1,0) + k3 (0,0,1) = (0,0,0)
(k1,0,0) + (0, k2,0) + (0,0,k1) = (k1, k2, k3) = (0,0,0)

Thus i, j and k are linearly independent

Similarly we can prove that the vectors

e1 = (1, 0, …, 0), e2 = (0, 1, 0, …, 0), …, en = (0, 0, …, 1)

are linearly independent

Example. Consider the vectors
v1=(1,–2,3), v2=(5,6,–1) and v3 = (3,2,1)
A solution k1, k2, k3 of the vector equation
k1v1 + k2v2 + k3v3 = 0
is also a solution of the homogeneous linear system
k1 + 5k2 + 3k3 = 0
–2k1 + 6k2 + 2k3 = 0
3k1 – k2 + k3 = 0
Since the determinant of the corresponding coefficient
matrix is 0, the system has a nontrivial solution.
Therefore
v1, v2 and v3 are linearly dependent
Example. Show that the polynomials p0=1, p1= x, p2=
x2, …, pn= xn are linearly independent in Pn

Solution. Consider a solution of the vector equation

a0p0 + a1p1 + a2p2 + … + anpn= 0
or equivalently
a0 + a1x + a2x2 + … + an xn = 0 for all x
Hence the polynomial
p(x) = a0 + a1x + a2x2 + … + an xn
has an infinite number of roots.
Therefore it must be the zero polynomial, i.e.
a0 = a1 = a2 = … = an = 0
Thus p0, p1, p2, …, pn are linearly independent
Theorem. A set S with two or more vectors is
(a) linearly dependent if and only if at least one of the
vectors in S is expressible as a (nontrivial) linear
combination of the other vectors in S
(b) linearly independent if and only if no vector in S
is expressible as a (nontrivial) linear combination
of the other vectors in S

Proof. (a) and (b) are clearly equivalent. Therefore we

need to prove (a) only.
Assume that
k1v1 + k2v2 + … + krvr = 0
has a nontrivial solution, where S = {v1, v2, …, vr}
Then at least one of the ki's is non zero, say k1  0.
We have
 k2   k3   kn 
v1 =  −  v 2 +  −  v 3 +  +  −  v n
 k1   k1   k1 

Conversely, assume that v1 is expressible as a linear

combination of v2, v3, …, vr,

v1 = c2v2 + … + crvr
Then
v1 – c2v2 – … – crvr = 0

Thus S is a linearly dependent set

Example. We have proved that the vectors
v1= (2,–1,0,3), v2= (1,2,5,–1) and v3 = (7,–1,5,8)
are linearly dependent by obtaining the relation
3v1 + v2 –v3 = 0
We can also prove this fact by observing that each of
these vectors is expressible as a linear combination of
the other two vectors
v1 = − 13 v 2 + 13 v 3 , v 2 = −3v1 + v 3 , v 3 = 3v1 + v 2

Example. We can prove that the standard basis

vectors i, j, k of R3 are linearly independent by
applying part (b) of the Theorem
Let assume indeed that k is expressible as a linear
combination of i and j. Then

k = k1i + k2j
(0,0,1) = k1(1,0,0) + k2 (0,1,0)
(0,0,1) = (k1, k2 , 0)

This equation has no solution so that k cannot be

expressible as a linear combination of i and j.

Similarly we see that i cannot be expressible as a linear

combination of j and k; and j cannot be expressible as a
linear combination of i and k.
Therefore i, j, and k are linearly independent
Theorem.
(a) A finite set of vectors that contains the zero vector is
linearly dependent
(b) A set with exactly two vectors is linearly independent
if and only if neither vector is a scalar multiple of the
other

Proof. (a) Let assume S = {v1, v2, …, vr , 0}. Then 0 is

expressible as a (nontrivial) linear combination of the
vectors in S
0v1 + 0v2 + … + 0vr + 1(0)= 0

Hence S is a linearly dependent set

Proof. (b) Let assume S = {v1, v2} and 0 is expressible
as a nontrivial linear combination

k1v1 + k2v2 = 0

Now k1 or k2 is nonzero. Let's assume k1  0

Then  k2 
v1 =  −  v 2
 k1 

Example. The following two functions satisfy (b)

so that they form an independent set
f1 = x and f2 = sinx
5.3.2. Geometric Interpretation of
Linear Independence
In R2 or R3:
❖ Two vectors are linear independent if and only if
they do not lie on the same line when they are placed
with their initial points at the origin
❖ Three vectors are linear independent if and only if
they do not lie on the same plane when they are
placed with their initial points at the origin k1v1+ k2v2

v2
v2
v1
v1
Theorem. Let S = {v1, v2, …, vr } be a set of vectors
in Rn. If r > n, then S is linearly dependent

Proof. Let v1 = (v11, v12, … , v1n)

v2 = (v21, v22, … , v2n)
… … …
vr = (vr1, vr2, … , vrn)

Now the vector equation

k1v1 + k2v2 + … + krvr = 0

has a nontrivial solution if and only if the system in

component form has a nontrivial solution
 v11k1+ v21k2+ … + vr1kr = 0
v12k1+ v22k2+ … + vr2kr = 0
… … …
v1nk1+ v2nk2+ … + vrnkr = 0

This homogeneous system of linear equations has

more unknowns than equations so that it has a
nontrivial solution

Therefore S = {v1, v2, …, vr } is linearly dependent

 5.3.3. Linear Independence of
Functions
To verify if n differentiable functions f1 = f1(x), f2 =
f2(x), …, fn = fn(x) are linearly independent, we form

 f1 ( x) f 2 ( x)  f n ( x) 
 f ' ( x) f 2 ' ( x)  f n ' ( x) 
 1
    
 ( n −1) ( n −1) ( n −1) 
 f1 ( x) f2 ( x)  fn ( x)

The determinant of which is called the Wronskian

of f1, f2, …, fn, and denoted by W(x)
Suppose that f1, f2, …, fn are linearly dependent
Then there are non simultaneously zero scalars k1,
k2, …, kn such that
k1f1(x) + k2 f2(x) + … + kn fn(x) = 0

differentiate this relation up to order n –1 we get the

linear system

k1 f1 ( x ) + k2 f 2 ( x ) +  + kn f n ( x ) =0
k1 f1 ' ( x ) + k2 f 2 ' ( x ) +  + kn f n ' ( x ) =0
   
( n −1) ( n −1) ( n −1)
k1 f1 ( x ) + k2 f 2 ( x ) +  + kn f n ( x ) = 0
Thus if f1, f2, …, fn are linearly dependent, then the
above homogeneous linear system has a nontrivial
solution, or equivalently, the matrix

 f1 ( x) f 2 ( x)  f n ( x) 
 f ' ( x) f 2 ' ( x)  f n ' ( x)  
 1
    
 ( n −1) ( n −1) ( n −1) 
 f1 ( x) f2 ( x)  fn ( x)

is not invertible
i.e. the Wronskian is identically 0 on (–, )
Thus we have proved

Theorem.
If the functions f1, f2, …, fn have continuous
derivatives up to order n –1 on the interval (–
, ) and if the Wronskian is not identically zero
on (–, ), then these functions form a linearly
independent set in C(n –1) (–, )

Note. The converse is false: if the Wronskian of

f1, f2, …, fn is identically zero on (–, ), then
these functions may be linearly independent
 Example.
Consider the functions f1, f2 on (–, ) defined by:

 0 1 if x  0
  0 1 if x  0

f1 ( x) =  − 2 f 2 ( x) =  − 2

e x
if x  0 
e x
if x  0

The Wronskian of f1 and f2 is identically zero on (–, ):

f1 f2  0 if x  0 since the 1st column = 0
W ( x) = =
f1 ' f 2 ' 0 if x  0 since the 2 nd column = 0

However f1 and f2 are clearly linearly independent in C1

(–, )
Example. Prove that the functions f1 = x and f2 =
sinx form a linearly independent set in C1 (–, )

Proof. The Wronskian of these functions is

x sin x
W ( x) = = x cos x − sin x
1 cos x

This function is not identically zero on (–, ), so

that the functions f1 = x and f2 = sinx are linearly
independent
Example. Prove that the functions f1 = 1, f2 = ex
and f3 = e2x form a linearly independent set in C2
(–, )

Proof. The Wronskian of these functions is

1 ex e2 x
W ( x) = 0 ex 2e 2 x = 2e 3 x
0 ex 4e 2 x
This function is not identically zero on (–, ).
Therefore the functions f1, f2 and f3 form a linearly
independent set in C2 (–, )
 5.4. BASIS AND DIMENSION

5.4.1. Non Rectangular Coordinate Systems

5.4.2. Coordinates Relative to a Basis
5.4.3. Dimension
5.4.4. Some Fundamental Theorems
5.4.1. Non Rectangular
Coordinate Systems
Recall that the coordinates (a,b) of a point P in the
plane with respect to a rectangular coordinate system
establishes a one-to-one correspondence between the
points in the plane and the pair of real numbers

y
b (a, b)

x
a
We can also use a non-rectangular coordinate system
consisting of two non-colinear axes to establish a one-
to-one correspondence between the points in the plane
and the pairs of real numbers
Let v1 and v2 be two unit vectors along these axes.
Then for any point P in the plane, we write the vector
OP as a linear combination of v1 and v2

OP = a v1 + b v 2 y
P(a, b)
by projecting parallel to the axes. b
v2
x
Then a, b are the coordinates of P
O v1 a
 Similarly a non-rectangular coordinate system in the 3-
space, consisting of three non-coplanar axes, establishes
a one-to-one correspondence between the points in space
and the ordered triple of real numbers

Let v1, v2 and v3 be three unit vectors along these axes,

then for any point P in the plane, we write the vector OP
as a linear combination of v1 , v2 and v3
z
by projecting parallel to the axes. c
P(a, b, c)
OP = a v1 + b v 2 + c v 3
0 y
Then a, b and c are the
a b
coordinates of P
x
Definition. Let V be any vector space and S = {v1, v2, …,
vn} be a set of vectors in V, then S is called a basis for
V if the following conditions hold
(a) S is linearly independent
(b) S spans V

Theorem. If S = {v1, v2, …, vn} is a basis for the vector

space V, then every vector v in V can be expressed in
a unique way in the form
v = c1v1 + c2v2 + … + cnvn

Proof. Since S spans V, every vector in V can be expressed

as a linear combination of v1, v2, …, vn
Consider two representations of a vector v in V

v = c1v1 + c2v2 + … + cnvn

v = k1v1 + k2v2 + … + knvn
We have
0 = (c1 – k1)v1 + (c2 – k2)v2 + … + (cn – kn)vn

Since S is a linearly independent set, we must have

c1 – k1 = 0, c2 – k2 = 0, …, cn – kn = 0
i.e.
c1 = k1, c2 = k2, … , cn = kn
Theorem. If S = {v1, v2, …, vm} is a set of vectors in
the vector space V such that S spans V, then S
contains a subset that is a basis for V

Proof. If S is linearly independent, then it is a basis

for V. Otherwise 0 is expressed as a nontrivial linear
combination of v1, v2, …, vm
0 = k1v1 + k2v2 + … + kmvm
We may assume that km  0, then vm is expressed as
a nontrivial linear combination of v1, v2, …, vm–1
Hence S1 = {v1, v2, …, vm –1} spans V. We can repeat
the process until we find a linearly independent subset
of S that spans V, i.e. a basis for V
5.4.2 Coordinates Relative to a Basis
If S={v1, v2, …, vn} is a basis for the vector space V, and
v = c1v1 + c2v2 + … + cnvn
is a representation of a vector v in V,

then the scalars c1, c2, …, cn are called the coordinates of

v relative to the basis S.
The corresponding vector (c1, c2, …, cn) in Rn is called
the coordinates vector of v relative to S and denoted by

(v)S = (c1, c2, …, cn)

Note that the coordinate vector depend on S and also in
the order of the vectors v1, v2, …, vn
Example. We have proved that the standard basis
vectors i, j, k of R3 are linearly independent
On the other hand, these vectors also span R3 since any
vector v in R3 can be expressible as
v = (a, b, c) = a(1,0,0) + b(0,1,0) + c(0,0,1)
= ai + bj + ck
S = {i, j, k} is called the standard basis of R3
We have
(v)S = (a, b, c) = v
 Example. We have seen that the vectors
e1 = (1, 0, 0, …, 0), e2 = (0, 1, 0, …, 0),
…, en = (0, 0, 0, …, 1)
form a linearly independent set S in Rn

On the other hand every vector v = (v1, v2, …, vn ) in

Rn can be expressed as a linear combination
v = v1e1 + v2e2 + … + vnen
Hence S is a basis for Rn called the standard basis for
Rn. The coordinate vector of v relative to S satisfy

(v)S = (v1, v2, …, vn ) = v

Example. Let v1 = (1,2,1), v2 = (2,9,0), v3 = (3,3,4)
Show that S = {v1, v2, v3} is a basis for R3
Solution. For every vector b = (b1, b2, b3 ) in R3 we
write
b = c1v1 + c2v2 + c3 v3

(b1, b2, b3 ) = c1 (1,2,1) + c2 (2,9,0) + c3 (3,3,4)

= (c1+ 2c2+ c3, 2c1+ 9c2 + 3c3, c1+ 4c3)

So (c1, c2, c3 ) is a solution of the system

c1 + 2c2 + 3c3 = b1
2c1 + 9c2 + 3c3 = b2
c1 + 4c3 = b3
 1 2 3
2 9 3 = −1
1 0 4
Therefore the above system has a solution (c1,c2,c3)
On the other hand, the following homogeneous linear
system has only the trivial solution
c1 + 2c2 + 3c3 = 0
2c1 + 9c2 + 3c3 = 0
c1 + 4c3 = 0
i.e. the vector equation c1v1 + c2v2 + c3 v3 = 0

has only the trivial solution c1 = c2 = c3 = 0

Hence S is a linearly independent set
Example. Let S = {v1, v2, v3} be the previous basis for R3.
(a) Find the coordinate vector of v=(5,–1,9) with respect
to S

Solution. We find c1, c2, c3 such that

v = c1v1 + c2v2 + c3 v3
(5,–1,9) = c1 (1,2,1) + c2 (2,9,0) + c3 (3,3,4)

So (c1, c2, c3 ) is a solution of the system

c1 + 2c2 + 3c3 = 5
2c1 + 9c2 + 3c3 = –1
c1 + 4c3 = 9
The system has a unique solution
c1 = 1, c2 = –1, c3 = 2
(v)S = (1, –1, 2)
(b) Find the vector v in R3 whose coordinate vector
with respect to S is (v)S = (–1,3,2)

Solution. We have
v = (–1)v1 + 3v2 + 2v3
= (–1)(1,2,1) + 3(2,9,0) + 2(3,3,4)
= (11, 31, 7)
Example. (a) Show that S = {1, x, x2, …, xn} is a basis
for Pn

Solution. We have shown that S is linearly independent

and spans Pn, i.e. S is a basis for Pn
We say that S is the standard basis for Pn
(b) Find the coordinate vector of p = a0 + a1x + a2x2
relative to the basis S = {1, x, x2} for P2
Solution. We have
p = a0 (1) + a1 (x) + a2 (x2)
Therefore
(p)S = (a0, a1, a2)
Example. Let S = {M1, M2, M3, M4} where

1 0 0 1  0 0  0 0 
M1 =   , M2 =   , M3 =   , M4 =  
 0 0   0 0  1 0   0 1 

a b
For an arbitrary 2  2 matrix M =  we have
c d 
a b  1 0 0 1 0 0 0 0 
M =  = a  + b  + c  + d 
c d  0 0 0 0 1 0  0 1 

M = aM1 + bM2 + cM3 + dM4

Hence S spans the vector space M22 of all 2  2 matrices

Moreover the relation
aM1 + bM2 + cM3 + dM4 = 0

may be written as
1 0 0 1 0 0 0 0  0 0 
a  + b  + c  + d  = 
 0 0   0 0   1 0   0 1   0 0 
 a b  0 0 
 c d  = 0 0 
   
Hence a = b = c = d = 0 and S is linearly independent

Thus S is a basis called the standard basis for M22.

The standard basis for Mmn is defined similarly
 5.4.3. DIMENSION
In R2 or R3:
❖ A line is a 1-dimensional subspace: it has a basis
consisting of a single nonzero vector
❖A plane is a 2-dimensional subspace: it has a basis
consisting of two non-colinear vectors
❖The whole R3 is a 3-dimensional subspace: it has a
basis consisting of three non-coplanar vectors k v + k v
1 1 2 2

v2 v2
v1
v1
 Definition. A vector space V is finite-dimensional if
it is the zero vector space or there exists a finite set
S = {v1, v2, …, vr } that forms a basis for V.
Otherwise, V is said to be infinite-dimensional

Example.
The vector spaces Rn, Pn , and Mmn are finite-dimensional.

The vector spaces F(–,), C(–,) and C(–,) are

infinite-dimensional.
Theorem.
Let V be a finite-dimensional vector space and
S = {v1, v2, …, vn } any basis.
(a) If a set has more than n vectors, then it is linearly
dependent
(b) If a set has fewer than n vectors, then it does not
span V

Corollary. All bases for a finite-dimensional vector

space have the same number of vectors
Definition. The dimension of a finite-dimensional
vector space V, denoted by dim(V), is defined as the
number of vectors in a basis for V.
The dimension of the zero vector space is 0

Note. It follows from the Corollary that dim(V)

does depend on the chosen basis for V

Example.
dim(Rn) = n
dim(Pn) = n + 1
dim(Mmn) = mn
Example. Determine a basis for and the dimension
of the solution space of the system

2x1 + 2x2 – x3 + x5 = 0
–x1 – x2 + 2x3 – 3x4 + x5 = 0
x1 + x2 – 2x3 – x5 = 0
x3 + x4 + x5 = 0
Solution.
The general solution of the system is
x1 = –s – t, x2 = s, x3 = –t, x4 = 0, x5 = t

Therefore, the solution vectors can be written as

  x1  − s − t  − s  − t  − 1 − 1
x   s   s   0  1 0
 2          
 x3  =  − t  =  0  + − t  = s  0  + t − 1
           
 x4   0   0   0  0 0
 x5   t   0   t   0   1 

− 1 − 1
Thus the vectors 1 0
   
v 1 =  0 , v 2 = − 1
   
0 0
 0   1 
span the solution space

And the dimension of the solution space is 2

5.4.4. Some Fundamental Theorems
Theorem. (Plus/Minus Theorem)
Let S be a nonempty subset in a vector space V.
(a) If S is linearly independent and v is a vector in V lying
outside of span(S) then S  {v} is still linearly
independent
(b) If v is a vector in S that is expressible as linear
combination of other vectors in S, then
span(S) = span(S –{v})
where S –{v} is the set obtained by removing v from S
Example. (a) A set S consisting of two non-colinear
vectors in R3 spans a plane through the origin.

If we add to S a vector v outside of this plane, then we

obtain three linearly independent vectors

(b) A set S consisting of three non-parallel vectors in R3

lying in a common plane through the origin will span the
plane.
If we remove from S any vector v that is a linear
combination of the other two, the remaining two vectors
still span the plane
Theorem. If V is an n-dimensional vector space, and if
S is a set in V with exactly n vectors, then S is a basis for
V if either S spans V or S is linearly independent
Example.

(a) The vectors v1 = (–3, 7) and v2 = (5, 5) are linearly

independent.
Therefore they form a basis for R2

(b) The vectors v1 = (2,0,–1), v2 = (4,0,7) form a

linearly independent set in the xz-plane, and the vector
v3 = (–1,1,4) lies outside of this plane so that v1, v2
and v3 are linearly independent.
Therefore they form a basis for R3
Theorem.
Let S be a finite set of vectors in a finite-dimensional
vector space V.
(a) If S spans V but is not a basis for V, then S can be
reduced to a basis for V by removing appropriate
vectors from S
(b) If S is a linearly independent set that is not a basis
for V, then S can be enlarged to a basis for V by
inserting appropriate vectors into S

Theorem. If W is a subspace of a finite-dimensional

vector space V, then dim(W)  dim(V)
Moreover if dim(W) = dim(V), then W = V
 5.5. ROW SPACES, COLUMN
SPACE & NULLSPACE
5.5.1. Column Space and Consistent Linear Systems
5.5.2. General & Particular Solutions
5.5.3. Bases for Row Spaces, Column Spaces, and
NullSpaces
5.5.4. Basis for Span(S)
5.5.1. Column Space and Consistent
Linear Systems
Definition. For an m  n matrix A = [aij], the vectors
r1 = [a11 a12 … a1n]
r2 = [a21 a22 … a2n]
………
rm = [am1 am2 … amn]
are called the row vectors of A, and the vectors
 a11   a12   a1n 
a  a  a 
c1 =  21 , c 2 =  22 ,  , c n =  2 n 
        
     
am1   am 2  amn 
are called the column vectors of A,
Example. Let 2 1 0
A= 
 3 − 1 4 

Then the row vectors of A are

r1 = [2 1 0] and r2 = [3 –1 4]

and the column vectors of A are

 2 1 0 
c1 =  , c 2 =  , c 3 =  
 3 − 1  4
Definition. If A is an m  n matrix, then
(a) The subspace of Rn spanned by the row vectors of A is
called the row space of A.
(b) The subspace of Rm spanned by the column vectors of
A is called the column space of A.
(c) The solution space of the homogeneous linear system
Ax = 0, which is a subspace of Rn, is called the
nullspace of A.

We will investigate:
❖ The relationship between the solutions of a linear system
Ax = b and the row space, column space and nullspace of A,
❖ and the relationship between these spaces
Let  a11 a12  a1n   x1 
a 
a22  a2 n   
x2 
A=  21
, and x = 
     
   
am1 am 2  amn   xn 
Let c1, c2, …, cn be the column vectors of A, then
Ax = x1c1 + x2c2 + … + xncn

Thus the linear system Ax = b is consistent if and

only if b is a linear combination of c1, c2, …, cn

Therefore we have
Theorem. A linear system Ax = b is consistent if and
only if b is in the column space of A

Example. Consider the linear system Ax = b

− 1 3 2   x1   1 
 1 2 − 3  x  =  − 9
  2   
 2 1 − 2  x3   − 3

Show that b is in the column space of A and express b

as a linear combination of the columns of A
Solution. Carry out Gaussian elimination we get
x1 = 2, x2 = –1, x3 = 3
 x1 = 2, x2 = –1, x3 = 3

Thus it follows from the Theorem that b is in the column

space of A and expressible as a linear combination of the
columns of A

1  −1  3   2 
 −9  = 2  1  −  2  + 3  −3
       
 −3  2  1   −2 
 5.5.2 General & Particular Solutions

Theorem. If x0 denotes a single solution of a consistent

linear system Ax = b, and if v1, v2, …, vk is a basis for the
nullspace of A, then every solution of Ax = b can be
expressed in the form
x = x0 + c1v1 + c2v2 + … + ckvk
Conversely, for all choice of the scalars c1, c2, …, ck, the
above vector x is a solution of Ax = b
Proof. Let x be an arbitrary solution, we have
A(x – x0) = Ax – Ax0 = b – b = 0
Hence x – x0 belong to the null space of A, so that
x – x0 = c1v1 + c2v2 + … + ckvk
i.e. x = x0 + c1v1 + c2v2 + … + ckvk
Conversely, it is obvious that the above vector x is a
solution of Ax = b

Note. The vector x0 is called a particular solution,

and the vector x is the general solution of Ax = b.
The expression c1v1 + c2v2 + … + ckvk is the general
solution of the homogeneous system Ax = 0.
 Example. Consider the linear system

x1 + 3x2 – 2x3 + 2x5 = 0

2x1 + 6x2 – 5x3 – 2x4 + 4x5 – 3x6 = –1
5x3 + 10x4 + 15x6 = 5
2x1 + 6x2 + 8x4 + 4x5 + 18x6 = 6

The general solution of the system is

x1 = –3r –4s –2t, x2 = r, x3 = –2s, x4 = s, x5 = t, x6 = 13

This result can be written as a linear combination of

column vectors
  x1  − 3r − 4 s − 2t  0 − 3 − 4 − 2
x   r  0   1   0   0 
 2          
 x3   − 2s  0   0   − 2   0 
 =  =   + r   + s  + t  
 x4   s  0   0   1   0 
 x5   t  0   0   0   1 
    1      
 x6   1
3   3   0   0   0 

The 1st column vector is a particular solution of the

non-homogeneous linear system.

The sum of the remaining column vectors is the general

solution of the corresponding homogeneous linear
system
 5.5.3. Bases for Row Spaces,
Column Spaces, and Nullspaces
Theorem. Elementary row operations do not change
the nullspace of a matrix

Example. A basis for the solution space of the system

2x1 + 2x2 – x3 + x5 = 0
–x1 – x2 + 2x3 – 3x4 + x5 = 0
x1 + x2 – 2x3 – x5 = 0
x3 + x4 + x5 = 0

consists of two column vectors

 − 1 − 1
1 0
   
v 1 =  0 , v 2 = − 1
   
0 0
 0   1 

This is also a basis for the nullspace of the matrix

2 2 −1 0 1
− 1 − 1 2 −3 1 
A=
1 1 −2 0 − 1
 
0 0 1 1 1
Theorem. Elementary row operations do not change
the row space of a matrix

Proof. Let B be obtained from A by performing an

elementary row operation T
✓ If T is a row interchange, then A and B have the
same row vectors, and hence the same row space
✓ If T is one of the other row operations, then the row
vectors of B are linear combinations of the row
vectors of A. Hence the row space of B is a subspace
of the row space of A.
Using T–1 we see that the row space of A is also a
subspace of the row space of B.
Note. Column space of a matrix may be changed by
elementary row operations.
For instance, let
 1 3
A= 
 2 6 
1 
Then the column space of A is span by c =  
2
If we add (–2) times the 1st row to the 2nd row,
then the matrix becomes
1 3
B=
0 0
1
The column space of B is span by c' =  
0 
Theorem. If a matrix R is in row-echelon form, then
(a) The nonzero row vectors of R form a basis for the
row space of R
(b) The pivot column vectors form a basis for the
column space of R

Proof. (a) The nonzero row vectors of R are clearly

independent and hence form a basis for the row space
of R
(a) The pivot column vectors are independent, and other
nonzero column vectors are linear combinations of
them.
Hence they form a basis for the column space of R
Example. Consider the matrix in row-echelon form
1 −2 5 0 3
0 1 3 0 
0
R=
0 0 0 1 0
 
0 0 0 0 0

The nonzero row vectors

r1 = [1 –2 5 0 3], r2 = [0 1 3 0 0], r3 = [0 0 0 1 0],

form a basis for the row space of R

The pivot column vectors
1  − 2 0 
0   1  0 
c1 =  , c 2 =  , c3 =  
0   0  1
     
0   0  0 
Are clearly independent, and
1  − 2  5   3 1
0   1   3 0  0 
11  + 5  =  , and   = 3 
0   0  0  0  0 
         
0   0  0  0  0 
Hence c1, c2, c3 form a basis for the column space of R
Theorem. If A and B are row equivalent matrices, then
(a) A given set of column vectors of A is linearly
independent iff the corresponding column vectors of B
are linearly independent
(b) A given set of column vectors of A forms a basis for
the column space of A iff the corresponding column
vectors of B form a basis for the column space of B
Example. Find bases for the row and column spaces of
 1 −3 4 −2 5 4 
 2 −6 9 −1 8 2 
A=
 2 −6 9 −1 9 7 
 
− 1 3 −4 2 −5 − 4

Solution. First reduce A to the row-echelon form

1 −3 4 −2 5 4 
0 0 1 3 −2 − 6
R=
0 0 0 0 1 5 
 
0 0 0 0 0 0 
The nonzero row vectors form a basis for the row space
of R, and hence a basis for the row space of A

r1 = [1 –3 4 – 2 5 4],
r2 = [0 0 1 3 –2 –6],
r3 = [0 0 0 0 1 5]
On the other hand the pivot column vectors form a
basis for the column space of R

1  4  5 
0  1   − 2
c1 ' =  , c 3 ' =  , c5 ' =  
0  0   1 
     
0  0   0 
Therefore the corresponding column vectors form a
basis for the column space of A

1  4   5 
2  9  8 
c1 =  , c 3 =  , c5 =  
2  9  9 
     
− 1  − 4 − 5
Example. Find a basis for the space spanned by the vectors

v1 = (1, –2, 0, 0, 3), v2 = (2, –5, –3, –2, 6),

v3 = (0, 5, 15, 10, 0), v4 = (2, 6, 18, 8, 6)

Solution. The space V spanned by these vectors is the

row space of the matrix

1 − 2 0 0 3
2 − 5 − 3 − 2 6
A=
0 5 15 10 0
 
2 6 18 8 6
 V is also the row space of the matrix

1 −3 0 0 3
0 1 3 2 0
 
0 0 1 1 0
 
0 0 0 0 0

A basis for V consists of the nonzero row vectors

w1 = (1,–3,0,0,3), w2 = (0,1,3,2,0), w3 = (0,0,1,1,0)

Note. The vectors of this basis are not row vectors of

the matrix A.
Example. Find a basis for the row space consisting of
row vectors of the matrix

1 − 2 0 0 3
2 − 5 − 3 − 2 6
A=
0 5 15 10 0
 
2 6 18 8 6

Solution. The remark suggests that we should change

to the column space of AT
We first reduce AT to the row-echelon form
 1 2 0 2 
0 1 −5 − 10
 
0 0 0 1 
 
0 0 0 0 

0 0 0 0 

The only pivot columns are the 1st, 2nd and the fourth.
Therefore a basis for the column space of AT consists of
the 1st, 2nd and fourth columns of AT
Passing to the transpose, we find a basis for the
row space of A consisting of the 1st, 2nd and fourth
rows of A
r1 = [1 –2 0 0 3], r2 = [2 –5 –3 –2 6],
r4 = [2 6 18 8 6]
 To find a subset of S  Rn that forms a basis for
span(S) and expresses the other vectors in S as linear
combinations of this basis we use
Procedure.
Step 1. Form the matrix A with column vectors, the vectors
in S.
Step 2. Reduce A to the row-echelon form R
Step 3. Identify the pivot columns of R. The corresponding
column vectors of A form a basis for span(S)
Step4. Express the other column vectors of R as linear
combinations of the pivot columns. The corresponding
equations for column vectors of A express S according
to this basis
Example. Find a subset of the set S of vectors
v1 = (1,–2, 0,3), v2 = (2,–5,–3,6), v3 = (0,1,3, 0), v4
= (2,–1,4,–7), v5 = (5,–8,1,2),
that forms a basis for span(S) and express other vectors in
S as linear combinations of this basis

Solution. We first construct the matrix with columns

the vectors of S

 1 2 0 2 5 
− 2 −5 1 −1 − 8
A=
 0 −3 3 4 1 
 
 3 6 0 −7 2
 Then reduce A to the row-echelon form
1 0 2 0 1
0 1 −1 0 1
R= 
0 0 0 1 1
 
0 0 0 0 0

The only pivot columns are the 1st, 2nd and the 4th.
Therefore a basis for the column space of A consists of
the 1st, 2nd and 4th columns of A
Hence {v1, v2, v4} is a basis for span(S)
On the other hand, we can express the columns w3,
w5 in terms of the columns w1, w2 , w4 of R
w3 = 2w1 – w2
w5 = w1 + w2 + w4
 w3 = 2w1 – w2
w5 = w1 + w2 + w4

The corresponding equations that express the

vectors v3, v5 as linear combinations of the vectors
v1, v2 , v4 are
v3 = 2v1 – v2
v5 = v1 + v2 + v4

These are called the dependency equations

5.6. RANK AND NULLITY

5.6.1. Rank and Nullity of a Matrix

5.6.2. Linear Systems and Rank
5.6.3. Overdetermined and
Underdetermined systems
5.6.1. Rank and Nullity of a Matrix
Given any matrix A, there are four fundamental matrix
spaces associated to A:
row space of A column space of A
nullspace of A nullspace of AT

Note that the row and column spaces of AT coincide with

the column and row spaces of A

Let R be a row-echelon form of A, then row spaces of A

and R are the same, a basis of which may be chosen
consisting of the nonzero rows of R
 dim(row space of A) = dim(row space of R)
= # nonzero rows of R

Now on each nonzero row of R, there exists a unique

pivot element that determines the corresponding pivot
column of R.
On the other hand the number of pivot columns of R is
also the number of elements in a basis for the column
space of A
Thus we have proved

Theorem. For any matrix A, the row space and the

column space of A have the same dimension
Definition.
▪ The common dimension of the row space and column
space of a matrix A is called the rank of A, and denoted
by rank(A).
▪ The dimension of the nullspace of A is called the
nullity of A and denoted by nullity(A)

Example. Find the rank and nullity of the matrix

− 1 2 0 4 5 − 3
 3 −7 2 0 1 4 
A= 
 2 −5 2 4 6 1
 
 4 −9 2 −4 −4 7 
Solution. The row-echelon form of A is
1 0 −4 − 28 − 37 13
0 1 −2 − 12 − 16 5 
R=
0 0 0 0 0 0
 
0 0 0 0 0 0

There are only two nonzero rows, and two pivot

columns (the first two), rank(A) = 2
To find the nullity we solve the linear system Ax = 0,
which may be written in the reduced form
x1 – 4x3 – 28x4 – 37x5 + 13x6 = 0
x2 – 2x3 – 12x4 – 16x5 + 5x6 = 0
Solving for the leading variables
x1 = 4x3 + 28x4 + 37x5 – 13x6
x2 = 2x3 + 12x4 + 16x5 – 5x6

Thus the general solution of the system is

x1 = 4r + 28s + 37t – 13u
x2 = 2r + 12s + 16t – 5u
x3 = r x4 = s
x5 = t x6 = u

or equivalently in terms of column vectors

  x1  4 28 37 − 13
 x  2 12  16   − 5 
 2        
 x3  1   0   0   0 
  = r   + s  + t   + u 
 x4   0   1   0   0 
 x5  0  0   1   0 
         
 x6  0  0   0   1 

Thus four vectors on the right form a basis for the

solution space. Therefore nullity(A) = 4
Theorem. For any matrix A,
rank(A) = rank(AT)

Proof.
rank(A) = dim(row space of A)
= dim(column space of AT) = rank(AT)

Lemma. For any m  n matrix A

(a) rank(A) = the number of leading variables in the
solution of Ax = 0
(b) nullity(A) = the number of parameters in the general
solution of Ax = 0
Proof. Let R be the reduced row-echelon form of A
(a) rank(A) = the number of pivot columns of R
= the number of leading variables in
the solution of Ax = 0
(b) nullity(A) = dim(solution space of Ax = 0)
= the number of parameters in
the general solution of Ax = 0
= the number of free variables
Since the number of leading variables + the number
of free variables = n we also have

Theorem. For any matrix A with n columns

rank(A) + nullity(A) = n
Example. Find the number of parameters in the general
solution of Ax = 0 if A is a 5  7 matrix of rank 3

Solution.
nullity(A) = n – rank(A) = 7 –3 = 4
Thus the number of parameters is 4

❖ Now let A be any m  n matrix A of rank r. Then

dim(row space of A) = dim(column space of A) = r

dim(nullspace of A) = n – r
dim(nullspace of AT) = m – r

Note that rank(A)  min(m,n)

 5.6.2. Linear Systems and Rank
Theorem. If Ax = b is a linear system of m equations
and n unknowns, then the following are equivalent
(a) Ax = b is consistent
(b) b is in the column space of A
(c) The coefficient matrix A and the augmented matrix
[A | b] have the same rank

Proof. b is in the column space of A if and only if A

and [A | b] have the same column space and hence
have the same rank
Example. Consider the linear system and its augmented
matrix
x1 – 2x2 – 3x3 + 2x4 = –4
–3x1 + 7x2 – 3x3 + x4 = –3
2x1 – 5x2 + 4x3 – 3x4 = 7
–3x1 + 6x2 + 9x3 – 6x4 = –1

1 −2 −3 2 − 4
− 3 7 −1 1 − 3
A=
 2 −5 4 −3 7 
 
− 3 6 9 −9 − 1
The row-echelon form of this matrix is

1 0 − 23 16 0
0 1 − 10 7 0

0 0 0 0 1
 
0 0 0 0 0

This system is inconsistent.

On the other hand the rank of the coefficient matrix
is 2 while the rank of the augmented matrix is 3.
Theorem. If Ax = b is a linear system of m equations in
n unknowns, then the following are equivalent
(a) Ax = b is consistent for every m  1 matrix b
(b) The column vectors of A span Rm
(c) rank(A) = m

Proof. (a) holds iff every vector b in Rm is also in

the column space of A. This is (b)
On the other hand (c) holds iff
dim(column space of A) = m ,
i.e the column space of A span Rm
5.6.3. Overdetermined &
Underdetermined Linear Systems
A linear system with more equations than unknowns
is called an overdetermined linear system

If Ax = b is an overdetermined linear system of m

equations in n unknowns (m>n), then the column
vectors of A cannot span Rm. It follows that

For A fixed, the overdetermined linear system Ax

= b cannot be consistent for all b in Rm
 Example. The overdetermined linear system

x1 – 2x2 = b1
x1 – x2 = b2
x1 + x2 = b3
x1 + 2x2 = b4
x1 + 3x2 = b5

cannot be consistent for all possible values of b1, b2,

b3, b4, b5.
We can find the relations satisfied by them so that the
corresponding linear system is consistent by using
Gauss-Jordan elimination
The reduced row-echelon form of the augmented
matrix is
1 0 2b 2 −b1 
0 1 b −b 
 2 1 
0 0 b 3 −3b 2 +2b1 
 
0 0 b 4 −4b 2 +3b1 
0 0 b 5 −5b 2 +4b1 

Thus the system is consistent if and only if

2b1 – 3b2 + b3 =0
3b1 – 4b2 + b4 =0
4b1 – 5b2 + b5 = 0
 2b1 – 3b2 + b3 =0
3b1 – 4b2 + b4 =0
4b1 – 5b2 + b5 = 0

Solving this homogeneous linear system we get

b1 = 5r – 4s,
b2 = 4r – 3s,
b3 = 2r – s,
b4 = r, b5 = s

where r and s are arbitrary

Note that if Ax = b is consistent, then a general
solution is the sum of a particular solution and the
general solution of the associated homogeneous
system Ax = 0.

Therefore the general solutions of the two systems

gave the same number of parameters,
and we have

Theorem. If Ax = b is a consistent linear system of

m equations in n unknowns, and if A has rank r,
then the general solution of the system contains n–r
parameters
Theorem. If A is an m  n matrix, then the following
are equivalent
(a) Ax = 0 has only the trivial solution
(b) The column vectors of A are linearly independent
(c) Ax = b has at most one solution for every m  1
matrix b

Proof. If Ax = b is consistent, then its general solution

and the general solution of Ax = 0 have the same
number of parameters. Hence (a)  (c)
Moreover this number of parameters is 0 iff
nullity(A) = 0  rank(A) = n
i.e. if and only if the column vectors of A are linearly
independent
A linear system with more unknowns than equations is
called an underdetermined linear system

If Ax = b is a consistent underdetermined linear system,

then its general solution has at least one parameter.
Hence the system has infinitely many solutions

Example. If A is a 5  7 matrix, then the system Ax = b

is underdetermined.
Let b be such that Ax = b is consistent,
then its general solution has 7 – r parameters, where r is
the rank of A
Now let A be an n  n matrix, then A is invertible if
and only if Ax = 0 has only the trivial solution

This means that A has nullity 0 or, equivalently, A has

rank n.

Of course this is the same as saying that the column

vectors of A are linearly independent or span Rn
(it follows that they form a basis for Rn)

By passing to the transpose, we find that this condition

is equivalent to the fact that the row vectors of A are
linearly independent or span Rn
(it follows that they form a basis for Rn)
Thus we have
Theorem. If A is an n  n matrix, then the following
are equivalent
(a) A is invertible
(b) Ax = 0 has only the trivial solution
(c) The column vectors of A are linearly independent
(d) The column vectors of A span Rn
(e) The column vectors of A form a basis for Rn
(f) The row vectors of A are linearly independent
(g) The row vectors of A span Rn
(h) The row vectors of A form a basis for Rn
(i) A has rank n
(j) A has nullity 0