LECTURE-4

Basis and Dimension

1
 Basis :
Basis
V: a vector space Linearly
Spanning Bases Independent
S = {v1, v2, …, vn}V Sets Sets

1) S spans V (i.e., span(S) = V) V

2) S is linearly independent

 S is called a basis for V

 Notes:
A basis S must have enough vectors to span V, but not so
many vectors that one of them could be written as a linear
combination of the other vectors in S
2
 Notes:
(1) the standard basis for R3:
{i, j, k}, for i = (1, 0, 0), j = (0, 1, 0), k = (0, 0, 1)
n
(2) the standard basis for R :
{e1, e2, …, en}, for e1=(1,0,…,0), e2=(0,1,…,0),…, en=(0,0,…,1)
Ex: For R4, {(1,0,0,0), (0,1,0,0), (0,0,1,0), (0,0,0,1)}

n
※Express any vector in R as the linear combination of the vectors in the
standard basis: the coefficient for each vector in the standard basis is the
value of the corresponding component of the examined vector,
e.g., (1, 3, 2) can be expressed as 1·(1, 0, 0) + 3·(0, 1, 0) + 2·(0, 0, 1)

3
(3) the standard basis for mn matrix space:
 aij  1
{ Eij | 1im , 1jn }, and in Eij 
other entries are zero
Ex: 2 2 matrix space:
1 0 0 1 0 0 0 0 
  ,  ,  , 
0 0   0 0   1 0   0 1 

(4) the standard basis for Pn(x):

{1, x, x2, …, xn}
Ex: P3(x) = {1, x, x2, x3}

4
 Ex 4.1: The nonstandard basis for R2
Show that S  {v1 ,v2 }  {(1,1), (1, 1)} is a basis for R2

 c1  c2  u1
(1) For any u  (u1 , u2 )  R , c1v1  c2 v 2  u  
2

c1  c2  u2
Because the coefficient matrix of this system has a nonzero determinant, the
system has a unique solution for each u. Thus you can conclude that S spans R2

c1  c2  0
(2) For c1 v1  c2 v 2  0  
c1  c2  0
Because the coefficient matrix of this system has a nonzero determinant, you
know that the system has only the trivial solution. Thus you can conclude that S is
linearly independent

5
 Theorem 4.1: Uniqueness of basis representation for any vectors
If S  v1 , v 2 ,, v n  is a basis for a vector space V, then every
vector in V can be written in one and only one way as a linear
combination of vectors in S
Pf:
 (1) span(S) = V
 S is a basis  
 (2) S is linearly independent
 span(S) = V Let v = c1v1+c2v2+…+cnvn
v = b1v1+b2v2+…+bnvn
 v + (–1)v = 0 = (c1–b1)v1 + (c2 – b2)v2 + … + (cn – bn)vn
S is linearly independent  with only trivial solution
 coefficients for vi are all zero
 c1 = b1 , c2 = b2 ,…, cn = bn (i.e., unique basis representation)6
Theorem 4.2: Bases and linear dependence
If S  v1 , v 2 ,, v n  is a basis for a vector space V, then
every set containing more than n vectors in V is linearly
dependent (In other words, every linearly independent set
contains at most n vectors)
Pf:
Let S1 = {u1, u2, …, um} , be another set such that m > n ,
we want to show that S1 is linearly dependent.
If V = span(S1)
u1  c11v1  c21v 2    cn1 v n
And uiV  u 2  c12 v1  c22 v 2    cn 2 v n

u m  c1m v1  c2 m v 2    cnm v n
7
Consider k1u1+k2u2+…+kmum= 0 (For L.I)
(if ki’s are not all zero, S1 is linearly dependent)
 d1v1+d2v2+…+dnvn= 0 (di = ci1k1+ci2k2+…+cimkm)

 S is L.I.  di=0 i i.e., c11k1  c12k 2    c1m k m  0

c21k1  c22k 2    c2 m k m  0

cn1k1  cn 2 k 2    cnmk m  0

 If the homogeneous system has fewer equations (n equations)

than variables (k1, k2, …, km), then it must have infinitely
many solutions
 m > n  k1u1+k2u2+…+kmum = 0 has nontrivial (nonzero) solution
 S1 is linearly dependent 8
 Theorem 4.3: Number of vectors in a basis
If a vector space V has one basis with n vectors, then every
basis for V has n vectors
Pf: ※ According to Thm. 4.2, every linearly independent set contains at most n vectors
of a vector space if there is a basis of n vectors spanning that vector space
S = {v1, v2, …, vn}
are two bases with different sizes for V
S' = {u1, u2, …, um}
S is a basis spanning V  
  m  n
S ' is a set of L.I. vectors  
n  m
S ' is a basis spanning V  
  n  m
S is a set of L.I. vectors  
※ For R3, since the standard basis {(1, 0, 0), (0, 1, 0), (0, 0, 1)} can span this vector
space, you can infer any basis that can span R3 must have exactly 3 vectors
9
 Dimension
 Dimension:
The dimension of a vector space V is defined to be the
number of vectors in a basis for V
V: a vector space S: a basis for V

 dim(V) = #(S) (the number of vectors in a basis S)

 Finite dimensional:
A vector space V is finite dimensional if it has a basis consisting
of a finite number of elements
 Infinite dimensional:
If a vector space V is not finite dimensional, then it is called
infinite dimensional
10
 Notes: dim(V) = n
Linearly
(1) dim({0}) = 0 Spanning
Bases Independent
Sets Sets
(If V consists of the zero vector alone,
the dimension of V is defined as zero)

(2) Given dim(V) = n, for SV #(S)  n #(S) = n #(S)  n

S: a spanning set  #(S)  n (Ex 3.2 on Slides 6 and 7 previous lec.)

S: a L.I. set  #(S)  n (from Theorem 4.2)
(Since a basis is defined to be a set of L.I.
S: a basis  #(S) = n vectors that can spans V, #(S) = dim(V) = n)
(see the above figure)

(3) Given dim(V) = n, if W is a subspace of V  dim(W)  n

※ For example, if V = R3, you can infer the dim(V) is 3, which is the number of
vectors in the standard basis
※ Considering W = R2, which is a subspace of R3, due to the number of vectors in
the standard basis, we know that the dim(W) is 2, that is smaller than 11
dim(V)=3
 Ex4.2: Find the dimension of a vector space according to
the standard basis
※ The simplest way to find the dimension of a vector space is to count the
number of vectors in the standard basis for that vector space

(1) Vector space Rn  standard basis {e1 , e2 ,  , en}

 dim(Rn) = n
(2) Vector space Mmn  standard basis {Eij | 1im , 1jn}
and in Eij  aij  1

other entries are zero
 dim(Mmn)=mn
(3) Vector space Pn(x)  standard basis {1, x, x2,  , xn}
 dim(Pn(x)) = n+1
(4) Vector space P(x)  standard basis {1, x, x2, }
12
 3
 Ex 4.3: Determining the dimension of a subspace of R
(a) W = {(d, c–d, c): c and d are real numbers}
(b) W = {(2b, b, 0): b is a real number}
(Hint: find a set of L.I. vectors that spans the subspace, i.e., find a basis for the
Sol: subspace.)
(a) (d, c– d, c) = c(0, 1, 1) + d(1, – 1, 0)
 S = {(0, 1, 1) , (1, – 1, 0)} (S is L.I. and S spans W)
 S is a basis for W
 dim(W) = #(S) = 2
(b) (2b, b,0)  b(2,1,0)
 S = {(2, 1, 0)} spans W and S is L.I.
 S is a basis for W
 dim(W) = #(S) = 1 13
 Ex 4.4: Finding the dimension of a subspace of M22
Let W be the subspace of all symmetric matrices in M22.
What is the dimension of W?
Sol:

a b  

W    a, b, c  R 

 b c  

a b  1 0 0 1 0 0
   a   b   c 
 b c  0 0   1 0   0 1 
1 0 0 1 0 0 
 S    ,  ,   spans W and S is L.I.
0 0 1 0 0 1 
 S is a basis for W  dim(W) = #(S) = 3
14
 Theorem 4.4: Methods to identify a basis in an n-dimensional
space
Let V be a vector space of dimension n
(1) If S  v1, v 2 , , v n is a linearly independent set of
vectors in V, then S is a basis for V
(2) If S  v1, v 2 , , v n  spans V, then S is a basis for V
(Both results are due to the fact that #(S) = n)
dim(V) = n

Linearly
Spanning
Bases Independent
Sets
Sets

#(S)  n #(S)  n 15
#(S) = n
Solve Problems

16