Math 415 - Lecture 9

Vector spaces and subspaces

Textbook: Chapter 2.1.

Suggested practice exercises: Chapter 2.1: 1, 2, 10, 11, 17, 18.

Khan Academy video: Linear Subspaces

1 Inverse of a matrix (cntd.)

We know how to find the inverse of a 2 × 2 matrix. What about the inverse 3 × 3 matrix or an n × n
matrix? Here we use the following Theorem from the last lectures.
Theorem 1. An n × n matrix A is invertible if and only if A is row equivalent to In , and in this
case, any sequence of elementary row operations that reduces A to In will also transform In to A−1 .

A few questions. Assume that A is invertible.

There are how many pivot positions in A?

By Theorem 1, In is an echelon form of A. Since In has n pivot positions, so does A.

How many free variables has the equation Ax = b?

Since A has n pivot positions and only n columns, every column of A must be a pivot column. Thus
there are no free variables.

Is there a b such that Ax = b is inconsistent?

Because A is row-equivalent of In , the augmented matrix [A|b] is row-equivalent to [In |c] for some c
in Rn . The latter matrix is of echelon form and can not a row of the form [0 . . . 0|d] where d 6= 0.

One important consequence of Theorem 1 is the following:

• A linear system Ax = b, where A is an n × n matrix has a unique solution if and only if A is
invertible.

Example 1. Use the Gauss Jordan method to compute the inverse of

 
1 −1 0
0 1 −1
−1 0 1

1
Solution.
 
1 −1 0 1 0 0
[A|I ]= 0 1 −1 0 1 0 
−1 0 1 0 0 1
 
1 −1 0 1 0 0
→  0 1 −1 0 1 0 
R3→R1+R3
0 −1 1 1 0 1
 
1 −1 0 1 0 0
→  0 1 −1 0 1 0 
R3→R2+R3
0 0 0 1 1 1
Failure: the reduced row echelon form of A will not be I, so A has no inverse!

2 Vector Spaces and Subspaces

• The most important property of column vectors in Rn is that you can take linear combinations
of them.

• There are many mathematical objects X, Y, . . . for which a linear combination cX + dY make
sense, and have the usual properties of linear combination in Rn
• We are going to define a vector space in general as a collection of objects for which linear
combinations make sense. The objects of such a set are called vectors.

Definition. A vector space is a non-empty set V of objects, called vectors, for which linear com-
binations make sense. More precisely: on V there are defined two operations, called addition and
multiplication by scalars (real numbers), subject to the ten axioms below. The axioms must hold for
all u, v, and w in V and for all scalars c and d.

1. u + v is in V . (V is “closed under addition”.)

2. u + v = v + u.
3. (u + v) + w = u + (v + w).
4. There is a vector (called the zero vector) 0 in V such that u + 0 = u.
5. For each u in V , there is a vector −u in V satisfying u + (−u) = 0.

6. cu is in V . (V is “closed under scalar multiplication”.)

7. c(u + v) = cu + cv.
8. (c + d)u = cu + du.

9. (cd)u = c(du).
10. 1u = u.

2
3 Vector Space Examples
  
a b
Example 2. Let M2x2 = : a, b, c, d ∈ R . Check that this is a vector space.
c d
Solution.

We need to say what the two operations are.

Addition:      
a b e f a+e b+f
+ = .
c d g h c+g d+h
Scalar Multiplication:    
a b ea eb
e· = .
c d ec ed
Next we need to say what the zero vector
 is. This is the matrix 0 such that 0 + A = A for any (2 × 2)
0 0
matrix A. Thus 0 vector must be . Then we need to check all the 10 axioms. They follow
0 0
from the corresponding properties of ordinary numbers.

Remark. 1. We can take instead of matrices of size 2 × 2 matrices of any shape: you can check
that the set Mm×n of m × n matrices is also a vector space, in the same way as we indicated
above.
2. Confusing: in the vector space M2×2 the vectors are in fact 2 × 2 matrices!
3. In the definition of the vector space M2×2 the multiplication of matrices plays no role; matrix
multiplication will show up when we study the connections between vector spaces.
 
  a
a b b
4. a “vector” behaves very much like a column vector  
 . A fancy person would say that
c d c
d
the vector spaces M2×2 and R4 are isomorphic.
Example 3. Let n ≥ 0 be an integer and let

Pn = the set of all polynomials of degree at most n.

Is this a vector space?
Solution.

3
 Yes, this is a vector space. Members of Pn have the form

p(t) = a0 + a1 t + · · · + an tn

where a0 , a1 , . . . , an are real numbers and t is a variable. We will just verify 3 out of the 10 axioms
here. Let p(t) = a0 + a1 t + · · · + an tn and q(t) = b0 + b1 t + · · · + bn tn and let c be a scalar.
The polynomial p + q is defined as follows:

(p + q)(t) = p(t) + q(t).

Therefore,

(p + q)(t) = p(t) + q(t)

= (a0 + b0 ) + (a1 + b1 )t + · · · + (an + bn )tn .

which is also a polynomial of degree at most n. So p + q is in Pn (i.e. Pn is closed under addition).

This verifies Axiom 1.
Next we need to find a zero vector. What this is the polynomial 0(t) such that 0(t) + p(t) = p(t)?
Take 0(t) = 0 + 0t + · · · + 0tn (zero vector in Pn ) Then

(p + 0)(t) = (a0 + a1 t + · · · + an tn ) + (0 + 0t + · · · + 0tn )

= (a0 + 0) + (a1 + 0)t + · · · + (an + 0)tn
= a0 + a1 t + · · · + an tn

and so p + 0 = p. This verifies Axiom 4.

Next we define scalar multiplication. Remember p(t) = a0 + a1 t + · · · + an tn . We define

(cp)(t) = cp(t) = (ca0 ) + (ca1 )t + · · · + (can )tn

which is in Pn . so that Axiom 6 holds.

The other 7 axioms also hold, so Pn is a vector space.

4 Subspaces
New vector spaces may be formed from subsets of other vector spaces. These are called subspaces.
Definition. A subspace of a vector space V is a subset H of V that satisfies 3 properties:

• The zero vector (of V ) belongs to H.

• If u, v both belong to H also the sum u + v belongs to H. (H is closed under vector addition).
• If u is in H and c is any scalar also cu belongs to H. (H is closed under scalar multiplication.)
Note that if the subset H satisfies these three properties, then H itself is a vector space.
 
0
Example 4. Z = is a subspace of R2 . Why?
0
Solution.

4
Check:
 
0
• is in Z.
0
     
0 0 0+0
• + = is in Z.
0 0 0+0
   
0 c0
• c = is in Z.
0 c0

Z is called the zero subspace of R2 . Every vectorspace has a zero subspace consisting just of the zero
vector.

 
1
Example 5. H = span is a subspace of R2 . Why?
1
Solution.
Check:
 
0
• is in H.
0
     
a b a+b
• + = is in H.
a b a+b
   
a ca
• c = is in H.
a ca

  
 a 
Example 6. Let H = 0 : a, b ∈ R . Show that H is a subspace of R3 .
b
 

Solution.

5
Verify properties 1, 2, and 3 of the definition of a subspace.
• The zero vector of R3 is in H.  
0
0 ∈ H, (a = b = 0)
0

• Adding two vectors in H always produces another vector whose second entry is 0 and therefore
the sum of two sectors in H is also in H. (H is closed under addition.)
     
a c a+c
 0  + 0 =  0  .
b d b+d

• Multiplying a vector in H by a scalar produces another vector in H. (H is closed under scalar

multiplication.)    
a ca
c 0 =  0  .
b cb

Since those three properties hold, H is a subspace of R3 .

Remark. Vectors (a, 0, b) look and act like the points (a, b) in R2 . But they are not the same!
  
x
Example 7. Is H = : x ∈ R a subspace of R2 ? (i.e. does H satisfy the properties of a
x+1
subspace?)

Solution.

H does not contain the zero vector (property 1).

   
x 0
=
x+1 0

cannot be true for any value of x.

Therefore, H is not a subspace!
Another way to show that H is not a subspace of R2 is to check whether H is closed under addition
(property 2).    
0 1
, ∈H
1 2
but    
0 1
+ ∈
/ H.
1 2

6
 x2
H

x1
−1 1 2 3

Example 8. Consider   
x
U= ∈ R2 : x 2 + y 2 < 1 .
y
Is this set U a subspace of R2 ?

x2

U
x1
−1 1 2

−1

Solution.

 
0.9
U is not a subspace. Note that U is not closed under scalar multiplication: While is in U ,
    0
0.9 1.8
2 = is not in U .
0 0

Example 9. Consider   
x 2
V = ∈ R : xy ≥ 0 .
y
Is this set V a subspace of R2 ?

7
 x2

x1
−1 1 2

−1

Solution.

   
0 −1
V is not a subspace. Note that U is not closed under addition: While and are in V , their
      1 0
0 −1 −1
sum + = is not in V .
1 0 1