Linear Algebra by Prof. R.

Vittal Rao

Lecture 3: August 12, 2010

3. (a) a × (b + c) = (a × b) + (a × c) ∀a, b, c ∈ Z
3.1 Integral Domain (b) (a + b) × c = (a × c) + (b × c) ∀a, b, c ∈ Z
(c) a, b ∈ Z, a , 0, b , 0 ⇒ a × b , 0 (that is a × b = 0 ⇒
Historically, computation with integers have been the ones used
either a or b (or both) must be 0.)
most often. These computations are facilitated by some funda-
mental properties of the structure of the integers, which we shall
now recall. Remark 3.1 The property (3c) is often used in solving polynomial
Let Z denote the set of all integers, equations.
Z = {0, ±1, ±2, . . . } .
When we generalize these notions we get the concept of Integral
With the integers, we perform two basic operations, namely, Domain.
+: addition and, ×: multiplication. Let S be any non-empty set (with at least two elements). Let
The basic properties of these two operations are listed below: + and × be two operations (binary operations) on S such that the
properties listed under 1, 2 and 3 hold with Z replaced by S. Then,
1. (a) ∀a, b ∈ Z ⇒ a + b ∈ Z
we say that (S, +, ×) is an Integral Domain.
(b) (a + b) + c = a + (b + c) ∀a, b, c ∈ Z Let us look at some examples.
(c) ∃0 ∈ Z such that a + 0 = a = 0 + a ∀a ∈ Z
(d) ∀a ∈ Z ∃(−a) ∈ Z such that a + (−a) = 0 = (−a) + a Example 3.1 Clearly, (Z, +, ×) is an integral domain.
(e) a + b = b + a ∀a, b ∈ Z

2. (a) a, b ∈ Z ⇒ a × b ∈ Z
Example 3.2 (Q, +, ×), (R, +, ×) and (C, +, ×) are integral domains.
(b) (a × b) × c = a × (b × c) ∀a, b, c ∈ Z
(c) ∃1 ∈ Z such that a × 1 = a = 1 × a

1
3.2 Field

Example 3.3 Let F2 = {0, 1} with + and × defined by the following

tables
3.2 Field
+ 0 1 × 0 1
0 0 1 0 0 0 Suppose now (S, +, ×) is an integral domain. For any a ∈ S we
1 1 0 1 0 1 have
Then, (F2 , +, ×) is an integral domain. 0 × a= (0 + 0) × a (since 0 + 0 = 0)
= (0 × a) + (0 × a) (Property (3b))
⇒ −(0 × a) + (0 × a)= −(0 × a) + ((0 × a) + (0 × a)) (Property (1d))
⇒ 0= (−(0 × a) + (0 × a)) + (0 × a) (Property (1d), (1b))
Example 3.4 In general, let p be any prime. Then, let 0= 0 + (0 × a) (Property (1d))
0= 0 × a (Property (2c))
Fp = 0, 1, 2, . . . , p − 1

Thus, we have, in any integral domain, (S, +, ×),
with
+: as ⊕p (addition modulo p), and 0 × a = 0 ∀a ∈ S.
×: as ⊗p (multiplication modulo p).
Hence, it does not make sense to ask for multiplicative inverse for 0.
Then, (Fp , ⊕p , ⊗p ) is an integral domain.
For an integral domain (S, +, ×), we say that a−1 ∈ S is multiplica-
tive inverse of a ∈ S if

a × a−1 = 1 = a−1 × a.
Example 3.5 Of p is not a prime then (Fp , ⊕p , ⊗p ) is not an integral
domain. Why? Note that as observed above 0−1 does not exist in any integral
domain. We observe that in the integral domains, (Q, +, ×),
(R, +, ×), (C, +, ×) every non-zero element has a multiplicative in-
verse, whereas in the integral domain (Z, +, ×) only 1 and −1 have
multiplicative inverses. This leads us to the notion of a field.
Example 3.6 Let R2×2 denote the set of all 2 × 2 real matrices. Let +
and × be the usual laws of addition and multiplication of matrices.
Then, (R2×2 , +, ×) is not an integral domain. Why? Definition 3.1 An integral domain in which every non-zero ele-
ment has a multiplicative inverse is called a field.

2
3.3 Re-look at diagonalization

1 1
!
Example 3.7 (Q, +, ×), (R, +, ×), (C, +, ×) are fields. then P−1
= 2
1
2 and
2
− 12
!
1 0
Example 3.8 (Z, +, ×) is not a field. P AP =
−1
.
0 −1

Since P and P−1 can be thought of as matrices (inverses of each

Example 3.9 When p is prime, (Fp , ⊕p , ⊗p ) is a field. other) over Q, R, C, it follows that A is diagonalizable over Q, R
and C. Let us now investigate diagonalizability over F2 . We have
! ! !
0 1 0 1 1 0
A =
2
= = I.
1 0 1 0 0 1
3.3 Re-look at diagonalization
Suppose P is a matrix over F2 which is invertible over F2 such that
Most of the computations we do will be over some field F. It is !
therefore necessary that we precisely state the field over which the d 1 0
P−1 AP = D = , a diagonal matrix
answers are being sought. We may get different types of answers 0 d2 !
based on the field over which we are seeking the answers. Let d21 0
⇒ P−1 APP−1 AP = D2 =
us illustrate this point with respect to diagonalization through the 0 ! d22
following examples. d2 0
⇒ P−1 A2 P = 1 2
0 d2 !
d2 0
Example 3.10 Consider the matrix ⇒ I = 1 2 since A2 = I
0 d2
! ⇒ d1 = 1, d2 = 1.
2 2
0 1
A= .
1 0 Since, P, P−1 , A are all matrices over F2 , we must have D as a matrix
over F2 and hence d1 , d2 ∈ F2 . So,
We can think of this as a matrix over
i) F2 or ii) Q or iii) R or iv) C. d1 = 1, d2 = 1
Let us look whether this is diagonalizable over any of these
! ⇒ D= I
1 1 ⇒ P−1 AP = D = I
fields. In the last lecture we saw that, if we take, P = ,
1 −1 ⇒ A= I

3
3.4 The Column Matrices

which is a contradiction since the given matrix A is not I. Thus, A

Example 3.12 We had seen in te last class that the matrix
is not diagonalizable over F2 whereas it is diagonalizable over Q,
R and C. !
0 −1
A=
1 0
Example 3.11 Consider the matrix,
! is diagonalizable over C but not over R.
0 2
A= .
1 0 These string of examples seem to suggest that working with C,
We can think of this as a matrix over i) Q or ii) R or iii) C. Suppose may be easier for diagonalization.
∃P over any of these fields such that
d1 0
!
3.4 The Column Matrices
P AP = D =
−1
, a diagonal matrix
0 d2 !
d21 0 While dealing with systems of equations,
⇒ P−1 APP−1 AP = D2 =
0 d22 Ax = b, A being an m × n matrix,
⇒ P−1 A2 P = D2
we have to deal with two types of vectors or column matrices, namely,
⇒ P−1 (2I)P = D2 since A2 = 2I
i) the given b, an m × 1 matrix, and
⇒ 2I = D2
ii) the unknown x, an n × 1 matrix.
⇒ d21 = 2 = d22 .
So, we shall look at the structure of column matrices.
Now we see that over Q, we cannot find d1 and d2 such that d21 = Let k be any positive integer, and F be any field. We denote,
2 = d22 . Hence, A is not diagonalizable over Q. However, if we take   
x



 
  1  


x
!   
1 1
  2
 
F = = , .
k
   
P = √1 − √1 x . : x ∈ 1 ≤ j ≤ k
   




 .
.


 j F 

2 2 
 
   

xk

 

as a matrix over R or C then over R or C,
 1 √ The addition + in F induces an addition + on Fk as follows:
 2 √12 

 x1 + y1 
! 
2 0
P−1 =  1  and P−1 AP = √ , a diagonal matrix.  x + y 
− √1  0 − 2
  2 2 
2 2 For x, y ∈ F , we define x + y = 
k
.  ∈ Fk .

 .
. 
Thus, A is diagonalizable over R and C but not over Q.
xk + y + k
 

4
3.5 Exercises

Multiplication in F induces a multiplication of an x ∈ Fk by an

α ∈ F. We call this scalar multiplication. We have for x ∈ Fk , α ∈ F,

αx1 
 
αx 
 2 
αx =  ..  .
 . 
αxn
 

Thus, on Fk we have two basic operations, namely,

1. addition, and 2. scalar multiplication.

3.5 Exercises
1. Let S = Z2 = {(a, b) : a, b ∈ Z} . Define + and × on S as follows:

(a, b) + (c, d) = (a + c, b + d)
(a, b) × (c, d) = (ac + 2bd, ad + bc)

(a) Examine whether (S, +, ×) is an integral domain. Find the

0 and 1 of this integral domain if it exists.
(b) Find an element x = (x1 , x2 ) ∈ S, if it exists, such that
x2 = 2.

2. Is (F12 = {0, 1, 2, 3, . . . , 11} , ⊕12 , ⊗12 ) an integral domain? If not,

which axioms of integral domain are not satisfied?

3. Find all solutions of x2 − 5x + 12 = 0 in

a) F2 , b) F8 , and, c) F12 .