ABSTRACT ALGEBRA II

Edited By
Richa Nandra
 Printed b y
EXCEL BOOKS PRIVATE LIMITED
A-45, Naraina, Phase-I,
New Delhi-110028
for
y
Lovel Professional University
Phagwara
 CONTENT

Unit 1: The Field of Quotient Euclidean Domains 1

Richa Nandra, Lovely Professional University

Unit 2: Principal Ideal Domains 9

Richa Nandra, Lovely Professional University

Unit 3: Unique Factorization Domains 20

Richa Nandra, Lovely Professional University

Unit 4: Polynomial Rings 26

Richa Nandra, Lovely Professional University

Unit 5: Division of Algorithm 35

Sachin Kaushal, Lovely Professional University

Unit 6: Irreducibility and Field Extensions 40

Sachin Kaushal, Lovely Professional University

Unit 7: Roots of a Polynomial 50

Richa Nandra, Lovely Professional University

Unit 8: Splitting Fields, Existence and Uniqueness 57

Richa Nandra, Lovely Professional University

Unit 9: Separable Extensions 62

Sachin Kaushal, Lovely Professional University

Unit 10: Galois Theory 76

Sachin Kaushal, Lovely Professional University

Unit 11: Computing Galois Groups 83

Richa Nandra, Lovely Professional University

Unit 12: Invariant Subfield 87

Richa Nandra, Lovely Professional University

Unit 13: The Galois Group of a Polynomial 92

Richa Nandra, Lovely Professional University

Unit 14: Solvability by Radicals 103

Richa Nandra, Lovely Professional University
 SYLLABUS
Abstract AlgebraII
Objectives:
 To learn about the structure as group, ring and field.
 To gain knowledge about homomorphisms, isomorphisms, cosets, quotient groups, and the isomorphism
theorems, rings, ideals, ring homeomorphisms, isomorphisms and its theorems.
 To learn about fields, quotient fields and field extensions Galois Theory also.

Sr. No. Content

1 Polynomial rings,The field of quotients Euclidean domains, Principal Ideal Domains,
Unique factorization domain
2 Prime fields, finite and algebraic extensions, Roots of a polynomial
3 splitting fields; existence and uniqueness, Separable extensions, Finite fields; the
structure, the existence of GF (pn)
4 Galois theory :Normal extensions, Galois groups
5 Symmetric functions, fundamental theorem, Constructible polygons, Solvability by
radicals
 Richa Nandra, Lovely Professional University

Notes Unit 1 : The Field of Quotient Euclidean Domains

CONTENTS
Objectives
Introduction
1 .1 Prime and Maximal Ideals
1 .2 Field of Quotients
1 .3 Summary
1 .4 Keywords
1 .5 Review Questions
1.6 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss whether an algebraic system is an integral domain or not
 Define and identify prime ideals and maximal ideals
 Prove and use simple properties of integral domains and fields
 Construct or identify the field of quotients of an integral domain

Introduction

Finally, we shall see how to construct the smallest field that contains a given
integral domain. This is essentially the way that Q is constructed from Z. We call
such a field the field of quotients of the corresponding integral domain.
In this unit, we have tried to introduce you to a lot of new concepts. You may need
some time to grasp them. Take as much time as you need. But by the time you
finish it, make sure that you have attained the knowledge of following topics.

1 .1 Prime and Maximal Ideals

In ‘Z’ we know that if p is a prime number and p divides the product of two integers
a and b, then either p divides a or p divides b. In other words, if ab  pZ, then
either a  pZ or b  pZ. Because of this property we say that pZ is a prime ideal, a
term we will define now.
Definition: A proper ideal P of a ring R is called a prime ideal of R if whenever ab 
P for a, b  R, then either a  P or b  P.
You can see that {0} is a prime ideal of Z because ab  {0}  a  {0} or b  {0}, where
a,b  Z.

LOVELY PROFESSIONAL UNIVERSITY

1
 Unit 1 : The Field of Quotient Euclidean
Domains

Another example of a prime ideal is Notes

Example: Let R be an integral domain. Show that I = ((0, x) | x  R) is a prime ideal of R x R.

Solution: Firstly, you know that I is an ideal of R x R. Next, it is a proper ideal since
I  R x R. Now, let us check if I is a prime ideal or not. For this let (a , b ), (a , b ) 
R x R such that (a , b )
1 1 2 2 1 1

(a , b )  I. Then (a a , b b ) = (0, x) for some x  R

2 2 1 2 1 2
 a a = 0, i.e., a, = 0 or a = 0, since R is a domain. Therefore, (a, b )  I or (a , b )  I. Thus, I is a
12 2 1 2 2
prime ideal.
Now we will, prove the relationship between integral domains and prime ideals.
Theorem 1: An ideal P of a ring R with identity is a prime ideal of R if and only if
the quotient ring R/P is an integral domain.
Proof: Let us first assume that P is a prime ideal of R. Since R has identity, so has
R/P. Now, let a + P and b + P be in R/P such that (ai – P) (b + P) = P, the zero
element of R/P. Then ab+P = P, i.e., ab  P. As P is a prime ideal of R either a  P
or b  P. So either a + P = P or b+P = P.
Thus, R/P has no zero divisors.
Hence, R/P is an integral domain.
Conversely, assume that R/P is an integral domain. Let a, b  R such that ab  P.
Then ab + P = P in R/P, i.e., (a + P) (b + P) = P in R/P. As R/P is an integral domain,
either a + P = P or b + P = P, i.e., either a E P or b  P. This shows that P is a prime
ideal of R.
An ideal mZ of Z is prime iff m is a prime number. Can we generalise this
relationship between prime numbers and prime ideals in Z to any integral domain?
To answer this let us first try and suitably generalise the concepts of divisibility and
prime elements.
Definition: In a ring R, we say that an elements divides an element b if b = ra for
some r  R. In this case we also say that a is a factor of b, or a is a divisor of b.
Thus, 3 divides 6 in Z , since 3.2  6.
7

Now let us see what a prime element is.

Definition: A non-zero element p of an integral domain R is called n prime element if
(i) p does not have a multiplicative inverse, and
(ii) whenever a, b  R and p | ab, then p | a or p | b.
Can you say what the prime elements of Z are? They are precisely the prime
numbers and their negatives.
Now that we know what a prime element is, let us see if we can relate prime ideals
and prime elements in an integral domain.
Theorem 2: Let R be an integral domain. A non-zero element p  R is n prime
element if and only if Rp is a prime ideal of R.
Proof: Let us first assume that p is a prime element in R. Since p does not have a
multiplicative inverse, 1  Rp. Thus, Rp is a proper ideal of R. Now let a, b  R such
that ab  Rp. Then ab = rp for some r  R
 p | a or p | b, since p is a prime element.
 a = xp or b = xp for some x  R.
a  p or b  Rp .
Thus ab  Rp  either a  Rp or b G  Rp, i.e., Rp is a prime ideal of R.

LOVELY PROFESSIONAL UNIVERSITY

2
 Notes

Note x R has a multiplicative inverse iff Rx = R.

Conversely, assume that Rp is a prime ideal. Then Rp  R, Thus, 1  Rp, and

hence, p does not have a multiplicative inverse. Now suppose p divides ab, where
a, b  R. Then ab = rp far some r  R, i.e., ab  Rp.
As Rp is a prime ideal, either a  Rp or b  Rp. Hence, either p | a or p | b. Thus, p
is a prime element in R.
Theorem 2 is very useful for checking whether an element is a prime element or
not, or for finding out when a principal ideal is a prime ideal.
Prime ideals have several useful properties.
Now consider the ideal 22 in Z. Suppose the ideal nZ in Z is such that 2Z  nZ Z. Then
n | 2.
 n=  1or n = 2.  nZ = Z or nZ = 2Z.
This shows that no ideal can lie between 2Z and Z. That is, 22 is maximal among
the proper ideals of Z that contain it. So we say that it is a “maximal ideal”. Let us
define this expression.
Definition: A proper ideal M of a ring R is called a maximal ideal if whenever I is an
ideal of R such that M  I  R, then either I = M or I = R.
Thus, a proper ideal M is a maximal ideal if there is no proper ideal of R which
contains it. An example that comes to mind immediately is the zero ideal in any
field F. This is maximal because you know that the only other ideal of F is F itself.
To generate more examples of maximal ideals, we can use the following
characterisation of such ideals.
Theorem 3: Let R be a ring with identity. An ideal M in R is maximal if and only if R/M is a field.
Proof: Let us first assume that M is a maximal ideal of R. We want to prove that R/M is a
field. For this, it is enough to prove that R/M has no non-zero proper ideals. So, let I be
an ideal of R/M. Consider the canonical homomorphism  : R  R/M :  (r) = r + M.
Then, you know that -1 (I) is an ideal of R containing M, the kernel of . Since M is a
maximal ideal of R. 1(I) = M or
-1(I) = R. Therefore, I = (-1 (I)) is either (M) or (R). That is, I = {0} or I = R/M,
where; = O +M = M. Thus, R/M is a field.
Conversely, let M be an ideal of R such that R/M is a field. Then the only ideals of R/M are
{0} and R/M. Let I be an ideal of R containing M. Then, as above (1) = {0} or, (I) =
R/M.

 I = -1((1)) is M or R. Therefore, M is a maximal ideal of R.

Corollary: Every maximal ideal of a ring with identity is a prime ideal.
Now, the corollary is a one-way statement. What about the converse? That is, is
every prime ideal maximal? What about the zero ideal in Z? Since Z is a domain but
not a field and Z Z/{0}, Z/{0} is a domain but not a field. Thus. (0) is a prime ideal
but not a maximal ideal of Z.

Example: Show that an ideal mZ of Z is maximal iff m is a prime number.

Solution: You know that Z,,, is a field iff m is a prime number. You also know that Z/mZ =
Z.
m
Z | mZ is a field iff m is prime. Hence, mZ is maximal in Z iff m is a prime number.

LOVELY PROFESSIONAL UNIVERSITY

3
 Unit 1 : The Field of Quotient Euclidean
Domains

Notes
(0,4,8)
Example: Show that 2Z12 is a maximal ideal of Z12, whereas is not.
Solution: You know that Z12 = Z/12Z and 2Z12 = 2Z/12Z. We see that Z12/ 2Z12 =
(Z/12Z)/ (2Z/12Z) = Z2, which is a field. Therefore, 2Z12  {0,2,4,6,8,10} is maximal
in Z12.

Now {0, 4, 8} = 4Z12  2Z12  Z12 .

Therefore, {0, 4, 8} is not maximal in Z .

12

We first introduced you to a special ideal of a ring, called a prime ideal. Its
speciality lies in the fact that the quotient ring corresponding to it is an integral
domain.
Then we discussed a special kind of prime ideal, i.e., a maximal ideal.

1.2 Field of Quotients

a
Consider Z and Q. You know that every element of Q is of the form , where a  Z and b  Z*.
b
a
Actually, we can also denote by the ordered pair (a, b)  Z × Z*. Now, in Q we
know that b
a c
 = - iff ad = bc. Let us put a similar relation on the elements of Z × Z*
b d

Now, we also know that the operations on Q are given by

a c ad  bc ac ac ac

  bd and .   ,  Q.

b d bd bd bd
Keeping these in mind we can define operations on Z × Z*. Then we can suitably
define an equivalence relation on Z × Z* to get a field isomorphic to Q.
We can generalise this procedure to obtain a field from any integral domain. So,
take an integral domain R. Let K be the following set of ordered pairs:
K= {(a,b) ) a , b  R and b  0)
We define a relation ~ in K by
(a, b) ~ (c, d) if ad = bc.
We claim that ~ is an equivalence relation. Let us see if this is so.

(i) (a, b) ~ (a, b)  (a, b)  K, since R is commutative. Thus, ~ is reflexive.

(ii) Let (a, b), (c, d)  K such that (a, b) ~ (c, d). Then ad = bc, i.e., cb = da. Therefore, (c, d) ~ (a, b). Thus, ~ is
symmetric.
(iii) Finally, let (a,b), (c,d), ( u, v)  K such that (a,b) – (c,d) and (c,d) ~ (u,v ). Then ad = bc and cv = du. Therefore,
(ad) v = (bc)v = bdu, i.e., avd =bud. Thus, by the cancellation law for multiplication (which is valid for a domain), we
get av = bu, i.e., (a,b) – (u,v). Thus, – is transitive.
Hence, ~ is an equivalence relation.

LOVELY PROFESSIONAL UNIVERSITY

4
 Notes Let us denote the equivalence class that contains (a,b) by [a,b]. Thus,

[a,b] = {(c,d) | c,d  R, d  0 and ad = bc }.

Let F be the set of all equivalence classes of K with respect to ~.
Let us define + and in F as follows. (It might help you to keep in mind the rules for
adding and multiplying rational numbers.)
[a,b] + [c,d] = [ad+bc,bd] and
[a,b].[c,d] = [ac,bd].
.
Do you think + and are binary operations on F?

Note b # 0 and d # 0 in the integral domain R imply bd # 0. So, the right-hand

sides of the equations given above are well defined equivalence classes. Thus,
the sum and product of two elements in F is again an element in F.
We must make sure that these operations are well defined.
So, let [a,b] = [a’,b’] and [c,d] = [c’,d’]. We have to show that [a,b] + [c,d] = [a’,b’]
+ [c’,d’], i.e., [ad+bc,bd] = [a’d’+b’c’,b’d’].
Now, (ad+bc) b’d’ – (a’d’ + b’c’) bd
= ab’dd’, + cd’bb’ – a’bdd’ – cdbb’
= (ab’ – a’b)dd’ + (cd’ - c’d) bb’
= (0) dd’ + (0)bb’, since (a,b) - (a’, b’) and (c,d) ~ (c’,d’).
= 0
Hence, [ad+bc,bd] = [a’ d’ + b’c’,brd’], i.e., + is well defined.
Now, let us show that [a,b] . [c,d] = [a’,br] . [c’,d’],
i.e., [ac,bd] = [a’c’ b’d’].
Consider (ac) (b’d’) - (bd) (a’c’)
= ab’cd’ – ba’dc’ = ba’cd’ – ba’ cd’, since ab’ = bar and cd’ = dc’
= 0
Therefore, [ac,bd] = [a’c’,b’d’]. Hence,. is well defined.
We will now prove that F is a field.
(i) + is associative : For [a,b], [c,d], [u,v] E F, ([a,b] + [c,d]) + [u,v] = [ad+bc,bd] + [u,v]
= [(ad+bc)v + ubd, Wv]
= [adv + b(cv+ud), bdv]
= [a,b] + [cv+ud,dv]
= [a,b] + ([c,d]+ [u,v])

LOVELY PROFESSIONAL UNIVERSITY

5
 Unit 1 : The Field of Quotient Euclidean
Domains

(ii) + is commutative :.For [a,b], [c,d]  F, Notes

[a,b] + [c,d] = [ad+bc,bd] = [cb+da,db] = [c,d] + [a,b]
(iii) [0,1] is the additive identity for F : For [a,b]  F,
[0,1] + [a,b] = [0.b+l.a, l.b] = [a,b]
(iv) The additive inverse of [a,b]  F is [–a,b] :
[a,b] + [–a,b] = [ab-ab,b2] = [0,b2] = [0,1], since 0.1 = 0.b2.
We would like you to prove the rest of the requirements for F to be a field.
So we have put our heads together and proved that F is a field.
Now, let us define f : R  F : f(.a) = [a,1]. We want to show that f is a monomorphism.
Firstly, for a, b  R,
f(a+b) = [a+b,1] = [a,]] + [b,l].
= f(a) + f(b), and
Thus, f is a ring homomorphism.
Next, let a,b  R such that f(a) = f(b). Then [a,1] = [b,l], i.e., a = b. Therefore, f is 1–1.
Thus, f is a monomorphism.
So, Im f = (R) is a subring of F which is isomorphic to R.
As you know, isomorphic structures are algebraically identical.
So, we can identify R with f(R), and think of R as a subring of F. Now, any element
of F is of the form [a,b] = [a,1] [l,b] = [a,l] [b,l] -1 = f(a) f(b)-1, where b  0. Thus,
identifying x  R with f(x)  f(R), we can say that any element of F is of the form
ab-1, where a,b  R, b  0.
All that we have discussed adds up to the proof of the following theorem.
Theorem 4: Let R be an integral domain. Then R can be embedded in a field F such
that every element of F has the form ab-1 for a, b  R, b  0.
The field F whose existence we have just proved is called the field of quotients (or
the field of fraction) of R.
Thus, Q is the field of quotient of Z. What is the field of quotients of R? The
following theorem answers this question.
Theorem 5: Iff : R  K is a monomorphism of an integral domain R into a field K, then there
exists a monomorphism g : F  K : g([a,1]) = f(a), where F is the field of quotients of R.
It says that the-field of quotients of an integral domain is the smallest field
containing it. Thus, the field of quotients of any field is the field itself.
So, the field of quotients of R is R and of Z is Z , where p is a prime number.
p p

Self Assessment

1. An ideal P of a ring R is called a/an .................. ideal of R. If whenever ab  P for a, b  R, then either a  P or b
 P.
(a) prime ideal (b) odd ideal
(c) even ideal (d) integer ideal

LOVELY PROFESSIONAL UNIVERSITY

6
 An ideal P of a ring R with identity is a prime ideal of R. If and only if
Notes 2. the .................. R/P
is an integral domain.
(a) polynomial ring (b) subring
(c) quotient ring (d) ideal ring
3. If x  R, it has multiplicative inverse iff RX = ..................

(a) R (b)
(c) XR (d)
4. A proper ideal m of a ring R is called maximal ideal of whenever I is an ideal of R such
that m .................. I .................. R then either I = m or I = R.
(a) , (b) ,
(c) ,  (d) , 
5. If R be a ring with identity. An ideal M in R is maximal if and only if .................. is a
field.
(a) R.M (b) R/M

(c) M/R (d) R+M

1.3 Summary

 The characteristic of any domain or field is either zero or a prime number.

 The definition and examples of prime and maximal ideals.
 The proof and use of the fact that a proper ideal I of a ring R with identity is prime (or
maximal) iff R/I is an integral domain (or a field).
 Every maximal ideal is a prime ideal.
 An element p of an integral domain R is prime iff the principal ideal pR is a prime
ideal of R.
 Z, is a field iff n is a prime number.
 The construction of the field of quotients of an integral domain.

1.4 Keywords

Prime Ideal: A ideal P of a ring R is called a prime ideal of R if whenever ab  P for

a, b  R, then either a  P or b  P.
Proper Ideal: A proper ideal M of a ring R is called a maximal ideal if whenever I is
an ideal of R such that M  I  R, then either I = M or I = R.
Maximal Ideal: Every maximal ideal of a ring with identity is a prime ideal.

1.5 Review Questions

1. Let F be a field. Show that F, with the Euclidean valuation d defined by d(a) = 1  a
 F/{0}, is a Euclidean domain.
2. Let F be a field. Define the function
d : F(x)\{0}  N {0} : d(f(x)) = deg f(x).
Show that d is a Euclidean valuation on F[x], and hence, F[x] is a Euclidean domain.

LOVELY PROFESSIONAL UNIVERSITY

7
 Unit 1 : The Field of Quotient Euclidean
Domains

3. Find all the units in Notes

(a) Z (b) Z (c) Z/5Z (d) Z + IZ
6

4. Let R be an integral domain. Show that

(a) u is a unit in R iff u | 1.
(b) for a, b  R, a | b and b | a iff a and b are associates in R.
5. Which of the following polynomials is irreducible? Give reasons for your choice.

(a) x2 – 2x + 1  R[x] (b) x2 + x + 1  C[x]

3
(c) x – i  C[x] (d) x – 3x2 + 2x + 5  R[x].

Answers: Self Assessment

1. (a) 2. (c) 3. (a) 4. (c) 5. (b)

1.6 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

8
 Richa Nandra, Lovely Professional University

Notes Unit 2 : Principal Ideal Domains

CONTENTS
Objectives
Introduction
2.1 Euclidean Domain
2.2 Principal Ideal Domain (PID)
2.3 Summary
2.4 Keywords
2.5 Review Questions
2.6 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss the Principal Ideal Domains
 Describe theorem related to Principal Ideal Domains

Introduction

In the last unit, you have studied about field and integer domain. In this unit, you
will study about Principal Ideal Domains.

2 .1 Euclidean Domain

In earlier classes you have seen that Z and F[x] satisfy a division algorithm. There
are many other domains that have this property. Here we will introduce you to
them and discuss some of their properties. Let us start with a definition.
Definition: Let R be an integral domain. We say that a function d : R \ (0)  NU (0)
is a Euclidean valuation on R if the following conditions are satisfied:

(i) d(a)  d (ab)  a, b  R \ {0}, and

(ii) for any a, b  k, b  0 3 q, r  R such that a = bq+r, where r = 0 or d(r) < d(b).
And then R is called a Euclidean domain.
Thus, a domain on which we can define a Euclidean valuation is a Euclidean domain.
Let us consider an example.

Example: Show that Z is a Euclidean domain.

Solution: Define, d : Z  N {0} : d(n) = |n|.

LOVELY PROFESSIONAL UNIVERSITY

9
 Unit 2 : Principal Ideal Domains

Then, for any a,b  Z \ {O], Notes

d(ab)= |ab| = |a| |b|  |a| (since | b| for b  0)
= d(a).
i.e., d(a)  d(ab).
Further, the division algorithm in Z says that if a, b  Z, b  0, then 3 q, r 
Z such that a = bq + r, where r = 0 or 0 < |r| < |b|.
i.e:, a = bq+r, where r = 0 or d(r) < d(b).
Hence, d is a Euclidean valuation and Z is a Euclidean domain.
Let us now discuss some properties of Euclidean domains. The first property
involves the concept of units. So let us define this concept. Note that this definition
is valid for any integral domain.
Definition: Let R be an integral domain. An element a  R is called a unit (or an
invertible element) in R, if we can find an element b  R, such that ab = 1, i.e., if a
has a multiplicative inverse.
For example, both 1 and -1 are units in Z since 1.1 = 1 and (-1).(-1) = 1.

!
Caution
The difference between a unit in R and the unity in R. The unity is the identity
with respect to multiplication and is certainly a unit. But a ring can have other
units ton, as you have just seen in the case of Z.
Now, can we obtain all the units in a domain? You know that every non-zero
element in a field F is invertible. Thus, the set of units of F’ is F \ {0}. Let us look at
some examples.

Example: Obtain all the units in F[x], where F is a field.

Solution: Let f (x)  P[ x] be a unit, Then  g(x)  F[x] such that f(x) g(x) = 1. Therefore,
deg f(x) g(x)) = deg(1) = 0, i.e.,
deg f(x) + deg g(x) = 0
Since deg f(x) and deg g(x) are non-negative integers, this equation can hold only if deg f(x) = 0
=deg g(x). Thus, f(x) must be a non-zero constant, i.e., an element of F\ {0}. Thus, the units of F[x] are the non-zero
elements of F. That is, the units of F and F[x] coincide.

Example: Find all the units in R = a  b 5|a,b  Z . 

Solution: Let a  b 5 be a unit in R. Then there exists

c  d 5  R such that

a  b 5 c  d 5 =1

 (ac – 5bd) + (bc + ad) 5 = 1

 ac – 5bd = 1 and bc+ad = 0
 abc – 5b2d = b and bc+ad = 0
 a(–ad) – 5b2d = b, substituting bc = –ad.

LOVELY PROFESSIONAL UNIVERSITY

10
 Notes So, if b  0, then (a2 + 5b2) | b, which is not possible.
 b = 0.
Thus, the only units of R are the invertible elements of Z.
Theorem 1: Let R be a Euclidean domain with Euclidean valuation d. Then, for any
a  R \ {0}, d(a) = d(l) iff a is a unit in K.
Proof: Let us first assume that a  R\ {0] with d(a) = d(1).
By the division algorithm in R,  q, r  R such that 1 = aq+r,
where r = 0 or d(r) < d(a) = d(1).
Now, if r  0, d(r) = d(r.1)  d(1). Thus, d(r) < d(1) can’t happen.
Thus, the only possibility for r is r = 0,
Therefore, 1 = aq, so that a is a unit.
Conversely, assume that a is a unit in R. Let b  R such that ab = 1. Then d(a) 
d(ab) = d(1). But we know that d(a) = d(a.1)  d(1). So, we must have d(a) = d(1).
Using this theorem, we can immediately solve Example, since f(x) is a unit in F[x] iff
deg f(x) = deg (1 ) = 0.
Now let us look at the ideals of a Euclidean domain.
Theorem 2: Let R be a Euclidean domain with Euclidean valuation d. Then every
ideal I % of R is of the form I = Ra for some a  R.
Proof: If I = (01, then I = Ka, where a = 0. So let us assume that I  {0}. Then I\ {0} is
non-empty.
Consider the set {d(a) | a  I \{0}). The well ordering principle this set has a minimal
element.
Let this be d(b), where b e I \ {0}. We will show that I = Rb.
Since b  1 and I is an ideal of R,
Rb  I.
Now take any a  I. Since I  R and R is a Euclidean domain, we can find q, r  R such
that
a = bq + r, where r = 0 or d(r) < d(b).
Now, b  I  bq  I. Also, a  I. Therefore, r = a – bq  I.
But r = 0 or d(r) < d(b), The way we have chosen d(b), d(r) < d(b) is not possible.
Therefore, r = 0, and hence, a = bq  Rb.
Thus, I  Rb.
From (1) and (2) we get
I = Rb.
Thus, every ideal I of a Euclidean domain R with Euclidean valuation d is principal,
and is generated by a  I, where d(a) is a minimal element of the set {d(x) | x  I \
(0) }.

Tasks 1. Show that every ideal of F[x] is principal, where F is a field.

2. Using Z as an example, show that the set
3. S = (a  R\ (0) | d(a) > d(1) } (0) is not an ideal of the Euclidean domain R
with Euclidean valuation d.

LOVELY PROFESSIONAL UNIVERSITY

11
 Unit 2 : Principal Ideal Domains

Notes
2 .2 Principal Ideal Domain (PID)
In the previous section you have proved that every ideal of F[x] is principal, where F is a field.
There are several other integral domains, apart from Euclidean domains, which have this property.
We give such rings a very appropriate name.
Definition: We call an integral domain R a principal ideal domain (PID, in short) if
every ideal in R is a principal ideal.

Note Every Euclidean domain is a PID

Thus, Z is a PID. Can you think of another example of a PID? What about Q and
Q[x]? In fact, by Theorem 2 all Euclidean domains are PIDs. But, the converse is not
true. That is, every principal ideal domain is not a Euclidean domain.

b
For example, the ring of all complex numbers of the form a  1  i 19 , where a,
b  Z, is a 2
principal ideal domain, but not a Euclidean domain.
Now let us look at an example of an integral domain that is not a PID.

Example: Show that Z[x] is not a PID,

Solution: You know that Z[x] is a domain, since Z is one. We will show that all its
ideals are not principal. Consider the ideal of Z[x] generated by 2 and x, i.e., <
2,x>. We want to show that < 2, x >  <f(x)> for any f(x)  Z[X].
On the contrary, suppose that 3 f(x)  Z[x] such that < 2, x > = < f(x) >. Clearly, f(x)  0.
Also, 3 g(x), h(x)  Z[x] such that
2 = f(x) g(x) and x = f(x) h(x).
Thus, deg f(x) + deg g(x) = deg 2 = 0 ...(1)
and deg f(x)+deg h(x) = deg x = 1 ...(2)
(I) shows that deg f(x) = 0, i.e., f(x)  Z, say f(x) = n.
Then (2) shows that deg h(x) = 1. Let h(x) = ax+b with a,b  Z.
Then x =f(x) h(x) = n(ax+b).
Comparing the coefficients on either side of this equation, we see that na = 1 and
nb = 0. Thus, n is a unit in Z, that is, n = If I.
Therefore, 1  < f(x) > = < x,2 >. Thus, we can write
I = x (a +a x+ ...+a xr ) + 2(b +b x+ .... +b xs), where a ,b  Z  i = 0, l,.. ...., r and j = 0, 1,...,s.
0 1 r 0 1 s i j

Now, on comparing the constant term on either side we see that 1 = 2b . This can’t be true, since
0
2 is not invertible in Z. So we reach a contradiction.
Thus, < x,2 > is not a principal ideal.
Thus, Z[x] is not a P.I.D.

LOVELY PROFESSIONAL UNIVERSITY

12
 Notes We will now discuss some properties of divisibility in PIDs. If R is a ring and a,b  R, with a  0
then a divides b if there exists c  R such that b = ac.
Definition: Given two elements a and b in a ring. R, we say that c  R is a common divisor o
a
and b if c | a and c | b.
An element d  R is a greatest common divisor (g.c.d, in short) of a, b  R if
(i) d | a and d | b, and
(ii) for any common divisor c of a and b, c | d.
For example, in Z a g.c.d of 5 and 15 is 5 , and a g.c.d of 5 and 7 is 1.
We will show you that if the g.c.d of two elements exists, it is unique up to units, i.e., if d and
are two g.c.ds of a and b, then d=ud’ , for some unit u.
So now let us prove the following result.
Theorem 3: Let R be an integral domain and a, b  R. If a g.c.d of a and b exists, then it is
unique
up to units.
Proof: So, let d and d’ be two g.c.ds of a and b. Since d is a common divisor and d’ is a g.c.d,
we
get d | d’ . Similarly, we get d’|d. Thus, we see that d and d’ are associates in R. Thus, the g.c.d o
a and b is unique up to units.
Theorem 3 allows us to say the g.c.d instead of a g.c.d. We denote the g.c.d of a and b by
(a,b).
(This notation is also used for elements of R × R. But there should be no cause for confusion.
The
context will clarify what we are using the notation for.
How do we obtain the g.c.d of two elements in practice? How did we do it in Z? We looked at the
common factors of the two elements and their product turned out to be the required g.c.d.
We will use the same method in the following example.

Example: In Q[x] find the g.c.d of

p(x) = x2 + 3x – 10 and
q(x) = 6x2 – 10x – 4
Solution: By the quadratic formula, we know that the roots of p(x) are 2 and –5, and the roots
of
q(x) are 2 and –1/3.
Therefore, p(x) = (x – 2) (x + 5) and q(x) = 2(x – 2) (3x + 1).
The g.c.d of p(x) and q(x) is the product of the common factors of p(x) and q(x), which is (x –
2).
Let us consider the g.c.d of elements in a PD.
Theorem 4: Let R be a PID and a, b  R. Then (a, b) exists and is of the form ax + by for some
x,y
 R.
Proof: Consider the ideal <a, b>. Since R is a PID, this ideal must be principal also.
Let d  R such that <a, b> = <d>. We will show that the g.c,d of a and b is d.
Since a  <d>, d | a, Similarly, d | b.
Now suppose c  R such that c | a and c | b.
Since d E <a,b>, 3 x, y  R such that d = ax+by.
Since c | a and c | b, c | (ax+by), i.e., c | d.

LOVELY PROFESSIONAL UNIVERSITY

13
 Unit 2 : Principal Ideal Domains

Thus, we have shown that d = (a,b), and d= ax+by for some x.y  R. Notes
The fact that F[x] is a PID gives-us the following corollary to Theorem 4.
Corollary: Let F be a field. Then any two polynomials f(x) and g(x) in F[x] have a
g.c.d which is of the form a(x)f(x)+b(x)g(x) for some a(x), b(x)  F[x].

1 (x)
For example, (c), (x–1) = (x3 – 2x2 + 6x – 5) + (x2 – 2x + 1).
5 5

Now you can use Theorem 4 to prove the following exercise about relatively prime
elements in a PID, i.e., pairs of elements whose g.c.d is 1.
Let us now discuss a concept related to that of a prime element of a domain.
Definition: Let R be an integral domain. We say that an element x  R is irreducible if
(i) x is not a unit, and
(ii) if x = ab with a,b  R, then a is a unit or b is a unit.
Thus, an element is irreducible if it cannot be factored in a non-trivial way, i.e., its
only factors are its associates and the units in the ring.
So, for example, the irreducible elements of Z are the prime numbers and their
associates. This means that an element in Z is prime iff it is irreducible.
Another domain in which we can find several examples is F[x], where F is a field.
Let us look at the irreducible elements in R[x] and C[x], i.e., the irreducible
polynomials over R and C. Consider the following important theorem about
polynomials in C[x]. You have already come across this in the Linear Algebra
course.
Theorem 5 (Fundamental Theorem of Algebra): Any non-constant polynomial
in C[x] has a root in C.
Does this tell us anything about the irreducible polynomials over C? Yes. In fact, we
can also write it as:
Theorem 5: A polynomial is irreducible in C[x] iff it is linear.
A corollary to this result is:
Theorem 6: Any irreducible polynomial in R[x] has degree 1 or degree 2.
We will not prove these results here but we will use them often when discussing
polynomials over R or C. You can use them to solve the following exercise.
Let us now discuss the relationship between prime and irreducible elements in a PID.
Theorem 7: In a PID an element is prime iff it is irreducible.

Proof: Let R be a PID and x  R be irreducible. Let x | ab, where a, b  R. Suppose x I a.

Then (x,a) = 1, since the only factor of x is itself, up to units. Thus, xb, Thus, x is prime.

Task Let R be a domain and p  R be a prime element. Show that p is irreducible.

(Hint: Suppose p = ab. Then p | ab. If p | a, then show that b must be a unit.)

LOVELY PROFESSIONAL UNIVERSITY

14
 Notes Now, why do you think we have said that Theorem, 7 is true for a PID only? You can see that one
way is true for any domain. Is the other way true for any domain? That is, is every irreducible
element of a domain prime? You will get an answer to this question.

Example: Just now we will look at some uses of Theorem 7.

Theorem 7 allows us to give a lot of examples of prime elements of F[x]. For example, any
linear
polynomial over F is irreducible, and hence prime. In the next unit we will particularly consider
irreducibility (and hence primeness) over Q[x].
Now we would like to prove a further analogy between prime elements in a PID and prime
numbers, namely, a result analogous. For this we will first show g very interesting property of
the ideals of a PID. This property called the ascending chain condition, says that any
increasing
chain of ideals in a PID must stop after a finite number of steps.
Theorem 8: Let R be a PID and I ,I ,.. .. .. be an infinite sequence of ideals of R satisfying.
1 2

1  1 ....
1 2

Then 3 m  N such that I, = I = I = ....

m+1 m+2


Proof: Consider the set I = I, I ... = In . We will prove that I is an ideal of R.
2
m1

Firstly, I  , since I   and I  I.

1 1

Secondly, if a,b  I, then a  I, and b  I for some r,s  N.

s

Assume r  s. Then I  I,. Therefore, a, b  I,. Since I, is an ideal of R, a–b  I,  I. Thus,

s

a–b  I  a, b  I.
Finally, let x  R and a  I. Then a  I, for some r  N.
 xa  I,  I. Thus, whenever x  R and a  I, xa  I.
Thus, I is an ideal of R. Since R is a PID, I = <a> for some a  R. Since a  I, a  I, for some m
 N.
Then I  I,. But I,  I. So we. see that I = I .
m

Now, I,  I  Therefore, I, = I

m+1 m m+1

Similarly, I, = I and so on. Thus, Im = I= I = ...

m+2 m+1 m+2

Now, for a moment let us go back, where we discussed prime ideals. Over there we
said that an element p  R is prime iff < p > is a prime ideal of R. If R is a PID, we
shall use Theorem 7 to make a stronger statement.
Theorem 9: Let R be a PID. An ideal < a > is a maximal ideal of R iff a is a prime element
of R.
Proof: If < a > is a maximal ideal of R, then it is a prime ideal of R. Therefore, a is
a prime element of R.

Conversely, let a be prime and let I be an ideal of R such that < a > I. Since R is a PID, I = < b
> for some b  R. We will show that b is a unit in R. < b > = R, i.e., I = R.
Now, < a >  < b >  a = bc for some c  R. Since a is irreducible, either b is an
associate of a or b is a unit in R. But if b is an associate of a, then <b> = <a>, a
contradiction. Therefore, b is a unit in R. Therefore, I = R.

LOVELY PROFESSIONAL UNIVERSITY

15
 Unit 2 : Principal Ideal Domains

Thus, <a> is a maximal ideal of R. Notes

What Theorem 9 says is that the prime ideals and maximal Ideals coincide in a PID.
Now, take any integer n. Then we can have n = 0, or n = ± 1, or n has a prime
factor. This property of integers is true for the elements of any PID, as you will see
now.
Theorem 10: Let R be a PID and a be a non-zero non-invertible element of R. Then
there is some prime element p in R such that a.
Proof: If a is prime, take p = a. Otherwise, we can write a =albl, where neither a, nor b is an
1
associate of a. Then < a >  < a >. If a is prime, take p = a . Otherwise, we can write a = a b ,
1 1 1 1 22

where neither a nor b is an associate of a,. Then <a >  < a >. Continuing in this way we get
2 2 1 2

an increasing chain
<a>  <a >  <a >  ...
1 2

By Theorem 8, this chain stops with some < a, >. Then a, will be prime, since it
doesn’t have any non-trivial factors. Take p = a,, and the theorem is proved.
And now we are in a position to prove that any non-zero non-invertible element of a
PID can be uniquely written as a finite product of prime elements (i.e., irreducible
elements).
Theorem 11: Let Rt be a PID. Let a  R such that a  0 and a is not a unit. Then a = p ,p ....p , where
1 2r

p ,p .... p , are prime elements of R.

1 2 r

Proof: If a is a prime element, there is nothing to prove. If not, then P | a, for some prime p in
1 1

R, by Theorem 10. Let a = p a . If p a . If a is a prime, we are through. Otherwise P | a, for some

l l 1 1 1 2

prime p in R. Let a , = p a . Then a = p p a . If a is a prime, we are through. Otherwise we

2 1 2 2 1 2 2 2
continue the process. Note that since al is a non-trivial factor of a, <a>  <a >. Similarly, <a > 
1 1

< a >. So, as the process continues we get an increasing chain of ideals,
2

<a>  <a >  <a >  ...

1 2

in the PID R. Just as in the proof of Theorem 10, this chain ends at < a, > for some
m  N, and a, is irreducible.
Hence, the process stops after m steps, i.e., we can write a = p p ... p , where p is a prime element
1 2 m i

of R  i = 1, .... m.
1 2 n 1 2 m

Thus, any non-zero non-invertible element in a PID can be

factorised into a product of primes. What is interesting about this factorisation is
the following result that you have already proved for Z in Unit 1.
Theorem 12: Let R be a PID and a  0 be non-invertible in R. Let a = p p ....p = q q ....q , where
p and q are prime elements of R. Then n = m and each p is an associate of some q for 1  i | n,
i j i j

1  j | m.
Before going into the proof of this result, we ask you to prove a property of prime
elements that you will need in the proof.

Task Use induction on n to prove that if p is a prime element in an integral domain

R and if p | a a ... a, (where a , a ,.. .., a,  R), then p | a , for some i = 1, 2. .... n.
1 2 l 2 i

Now let us start the proof of Theorem 12.

LOVELY PROFESSIONAL UNIVERSITY

16
 Notes Proof: Since p p , ...p, = q q ... ,.q , p | q q . ... q ,.
1 2 1 2 m 1 1 2 m

| q for some j = 1. .... ..,m. By changing the order of the q , if necessary, we ca

Thus, p assum
1 j i

that j = 1, i.e., p | q. Let q = p u . Since q is irreducible, u must be a unit in R. So p and q ar

1 1 l l 1 1 1 1

associates. Now we have

pp = P (p u )q ....q .
1 2 n 1 1 2 m
Cancelling p from both sides, we get
1

p p ...p = u q ...,q .
2 3 n 1 2 m

Now, if m > n, we can apply the same process to p , p , and so on.

2 3

Then we will get

1 = u u .... u q .... q .
1 2 n n+1 m

This shows that q is a unit. But this contradicts the fact that q is irreducible.
n+1 n+1

Thus, m  n.
Interchanging the roles of the ps and qs and by using a similar argument, we get n  m.
Thus, n = m.
During the proof we have also shown that each pi is an associate of some q, and vice
versa.
j

What Theorem 12 says is that any two prime factorisations of an element in a PID
are identical, apart from the order in which the factors appear and apart from
replacement of the factors by their associates.
Thus, Theorems 11 and 12 say that every non-zero element in a PID R, which is not
a unit, can be expressed uniquely (up to associates) as a product of a finite number
of prime elements.
For example, x2 – 1  R[x] can be written as (x-1)(x+1) or (x-1) (x-1) or [2(x-tl)] [2(x-1)] in
R[x].
The property that we have shown for a PID in Theorems 11 and 12 is true for
several other domains also. Let us discuss such rings now.

Self Assessment

1. An integral domains R a .................. if every ideal in R is a principle ideal.

(a) principle ideal domain (b) unique ideal domain
(c) special ideal domain (d) range ideal domain
2. g.c.d. represent ..................
(a) greatest common divisor (b) greatest common dividend
(c) greatest common domain (d) greatest commutated domain
3. Let R be .................. and a, b  R, if a g.c.d. of a and b exists, then it is unique up to
unit.
(a) domain and range (b) integral domain
(c) UID (d) SID

4. PID stands for ..................

(a) principal integral domain (b) pair ideal domain
(c) principle ideal domain (d) principle ideal divisor

LOVELY PROFESSIONAL UNIVERSITY

17
 Unit 2 : Principal Ideal Domains

5. Any irreducible polynomials R[x] has degree 1 or define .................. Notes

(a) n (b) 2
(c) 4 (d) 5

2 .3 Summary

 The definition and examples of a Euclidean domain.

 Z any field and any polynomial ring over a field are Euclidean domains.
 Units, associates, factors, the g.c.d. of two elements, prime elements and irreducible elements in an integral
domain.
 The definition and examples of a principal ideal domain (PID).
 Every Euclidean domain is a PID, but the converse is not true. Thus, Z, F and F[x] are PIDs, for any field F.
 The g.c.d. of any two elements a and b in a PID R exists and is of the form ax + by for some x, y  R.
 The Fundamental Theorem of Algebra: Any non-constant polynomial over C has all its roots in C.

2 .4 Keywords

Euclidean Domain: An integral domain D is called a Euclidean domain if for

each non-zero element x in D there is assigned a non-negative integer (x) such
that
(i) (ab) (b) for all non-zero a,b in D, and
(ii) for any non-zero elements a,b in D there exist q,r in D such that a = bq + r, where either r = 0 or (r) < (b).
UID: Let R be a commutative ring with identity. A non-zero element p of R is said to
be irreducible if
(i) p is not a unit of R, and
(ii) if p = ab for a,b in R, then either a or b is a unit of R. Any principal ideal domain is a unique factorization
domain.

2 .5 Review Questions

1. Show that a subring of a PID need not be PID.

2. Will any quotient ring of a PID be a PID? Why? Remember that a PID must be an integral domain.
3. Let R be an integral domain. Show that
(a) u is a unit in R iff u | 1.
(b) for a, b  R, a | b and b | a iff a and b are associates in R.

LOVELY PROFESSIONAL UNIVERSITY

18
 Notes 4. Find the g.c.d. of
(a) (b) x2 + 8x + 15 and x2 + 12x + 35 in Z[x],
2 and 6 in Z/  8 ,

(c) x3 – 2x2 + 6x – 5 and x2 – 2x + 1 in Q[x].

5. Let R be a PID and a, b, c  R such that a (bc. Show that if (a, b) = 1, then a | c.
(Hint: By Theorem 4,  x, y  R such that ax + b = 1.)

Answers: Self Assessment

1. (a) 2. (a) 3. (b) 4. (c) 5. (b)

2 .6 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

19
Richa Nandra, Lovely Professional University Unit 3 : Unique Factorization Domains

Unit 3 : Unique Factorization Domains Notes

CONTENTS
Objectives
Introduction
3.1 Unique Factorisation Domain (UFD)
3.2 Summary
3.3 Keyword
3.4 Review Questions
3.5 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss unique factorization domains
 Explain theorems of UID

Introduction

In this unit, we shall look at special kinds of integral domains. These domains were
mainly studied with a view to develop number theory. Let us say a few introductory
sentences about them.
You saw that the division algorithm holds for F[x], where F is a field. You saw that it
holds for Z. Such integral domains are called Euclidean domains.
We shall look at some domains which are algebraically very similar to Z. These are
the principal ideal domains, so called because every ideal in them is principal.
Finally, we shall discuss domains in which every non-zero non-invertible element can be
uniquely factorised in a particular way. Such domains are very appropriately called
unique factorisation domains. While discussing them, we shall introduce you to
irreducible elements of a domain.
While going through the unit, you will also see the relationship between Euclidean
domains, principal ideal domains and unique factorisation domains.

3.1 Unique Factorisation Domain (UFD)

Here we shall look at some details of a class of domains that include PDs.
Definition: We call an integral domain R a unique factorisation domain (UFD,
in short) if every non-zero element of R which is not a unit in R can be uniquely
expressed as a product of a finite number of irreducible elements of R.

LOVELY PROFESSIONAL UNIVERSITY

20
 Notes Thus, if R is a UFD and a  R, with a  0 and a being non-invertible, then
(i) a can be written as a product of a finite number of irreducible elements, and
(ii) if a = p p .. ..p = q q ..... q , be two factorisations into irreducibles, then n = m and
1 2 n 1 2 m

p is an associate of some q, where 1 | i | n, 1  j  m.

i j

Thus, F[x] is a UFD for any field F.

Also, since any Euclidean domain is a PID, it is also a UFD. You directly proved that Z is a UFD.
Why don’t you go through that proof and then try and solve the exercises.
Now we give you an example of a domain which is not a UFD (and hence, neither a
PID nor a Euclidean domain).

ED implies PID implies UFD

Theorem 1: Every Euclidean domain is a principal ideal domain.

Proof: For any ideal l, take a nonzero element of minimal norm b. Then l must be
generated by b, because for any a  l we have a = bq + r for some q, r with N(r) <
N(b), and we must have r = 0 otherwise r would be a nonzero element of smaller
norm than b, which is a contradiction.
Fact: If R is a UFD then R[x] is also a UFD.
Theorem 2: Every principal ideal domain is a unique factorization domain.
Proof: We show it is impossible to find an infinite sequence a , a ,... such that a is divisible
by a
1 2 1 i+1

but is not an associate. Once done we can iteratively factor an element as we are
guaranteed this process terminates.
Suppose such a sequence exists. Then the a generate the sequence of distinct principal ideals
(a )
i 1
(a )  ... The union of these ideals is some principal ideal (a). So a  (a ) for some n, which
2 n
implies (a ) = (a ) for all i  n, a contradiction.
i n

Uniqueness: Each irreducible p generates a maximal ideal (p) because if (p)  (a)
 R then p = ab for some b  R implying that a or b is a unit, thus (a) = (p) or (a) =
R. Thus R/(p) is a field. Next suppose a member of R has two factorizations
p ... p = q ... q
1 r 1 s

Consider the ideals (p ) . 1 (q ). Relabel so that i p i generates a 1 minimal

ideal amongst these
i i 1
(in other words, (p ) does not strictly contain another one of the ideals). Now we show (p ) = (q )
1 1 i

for some i. Suppose not. Then (p ) does not contain any q , thus q is nonzero modulo (p ) for all i, which is a contradiction because the left-hand side of the above equation is zero modulo (p ).

Relabel so that (p ) = (q ). Then p = uq for some unit u. Cancelling gives up ... p

1 1 1 1 2 r2
element up is also irreducible, so by induction we have that factorization is unique.
2

The converse of the above theorem is not always true. Consider the ring [x]. The
ideal (2, x) is not principal: suppose (2, x) = (a) for some a. Since this ideal contains
the even integers, a must be some integer (multiplication never reduces the degree
of an element), and in fact it must be (an associate of) 2. But (2) does not contain
polynomials with odd coefficients, so (2, x) = (2).

Example: Show that Z[ 5]  {a  b 5|a,b  Z} is not a UFD.

Solution: Let us define a function

f : Z [ 5 ]  N U {0} by f(a+b 5 ) = a2+ 5b2.

LOVELY PROFESSIONAL UNIVERSITY

21
 Unit 3 : Unique Factorization Domains

This function is the norm function, and is usually denoted by N. Notes

You can check that this function has the property that

f() = f() f()  , ,  Z [ 5 ].

Now, 9 has two factorisations in Z[ 5], namely,

9=3.3= (2+ –5 )(2– –5 ).

You have already shown that the only units of Z [ 5 ] are 1 and –1. Thus, no two of
3, 2+ 5 and 2 – 5 are associates of each other.

Also, each of them is irreducible. For suppose any one of them, say 2 + 5 , is

reducible. Then 2+ 5 =  for some non-invertible a,  Z[ 5 ].

Applying the function f we see that

f(2+ –5 ) = f() f(),

i.e., 9 = f() f().

Since f(), f()  N and a,  are not units, the only possibilities are f() = 3 = f().

So, if a = a + b 5 , then a2 + 5b2 = 3.

But, if b  0, then a2 + 5b2  5; and if b = 0, then a2 = 3 is not possible in Z. So we reach a
contradiction. Therefore, our assumption that 2 + 5 is reducible is wrong. That is, 2
+ 5 is irreducible.

Similarly, we can show that 3 and 2– 5 are irreducible. Thus, the factorisation of 9 as
a product of irreducible elements is not unique. Therefore, Z[ 5 ] is not a UFD.
From this example you can also see that an irreducible element need not be a prime element.

For example, 2 + 5 is irreducible and 2+ 5 |3.3, but 2+ 5 | 3 . Thus, 2 + 5 is

not a prime element.
Now let us discuss some properties of a UFD. The first property says that any two
elements of a UFD have a g.c.d. and their g.c.d. is the product of all their common
factors. Here we will use the fact that any element a in a UFD R can be written as

a  p1r1 p2r2 ...pnrn

where the pis are distinct irreducible elements of R. For example, in Z[x]
we have x3 – x2 – x + 1 = (x – 1) (x + l) (x – 1 ) = (x – 1)2 (x + 1).
So, let us prove the following result.
Theorem 3: Any two elements of a UFD have a g.c.d.
Proof: Let R be a UFD and a.b  R.

Let a  p1r1 p2r2 ...pnrn and b  p1s1 p2s2 ...pnsn

LOVELY PROFESSIONAL UNIVERSITY

22
 where p , p , ..., p are distinct irreducible elements of R and r and s are non-negative integers
Notes 
1 2 n i i

i = 12, ..., n.
(If some pi does not occur in the factorisation of a, then the corresponding r i = 0. Similarly, if
some p is not a factor of b, then the corresponding s = 0. For example, take 20 and 15 in Z.
Then
i i

20 = 22 × 30 × 5 and 15 = 20 × 31 × 51.)
Now, let t, = min (r , s )  i = 1, 2 ,....,n .
i i
divides a as well as b, since t  r and t 
t t
Then d  p 1 p 2 ...p tn s  i = 1, 2, ...., n.
1 2 n ii i i

Now, let c | a and c | b. Then every irreducible factor of c must be an irreducible

factor of a and of b, because of the unique factorisation property.
m m
Thus, c  p 1 p m2 ...p n , where m  r and m  s  i = 1,2, ...,n . Thus, m  t  i = 1,2 ,
1 2 n ii ii ii

Therefore, c | d.
Hence, d = (a, b).
Over there we found a non-UFD in which an irreducible element need not be a
prime element. The following result says that this distinction between irreducible
and prime elements can only occur in a domain that is not a UFD.
Theorem 4: Let R be a UFD. An element of R is prime iff it is irreducible.
Proof: We know that every prime in R is irreducible. So let us prove the converse.
Let a  R be irreducible and let a | bc, where b, c  R.
Consider (a,b). Since a is irreducible, (a,b) = 1 or (a,b) = a.
If (a,b) = a, a| b.

If (a,b) = I, then d| b . Let bc = ad, where d  R.

Let b  p1r1 p2r2 ...pmrm , and c  q1s1 q2s2 ...qnsn , be irreducible factorisations of b and c. Since bc
= ad
and a is irreducible, a must be one of the pis or one of the qs. Since a| b, a  p for any i.
j i

Therefore, a = q for some j. That is, a | c.

j

Thus, if (a,b) = 1 then alc.

So, we have shown that a/ bc  a | b or a | c.
Hence, a is prime.
Theorem 5: Let R be a UFD. Then R[x] is a UFD.
We will not prove this result here, even though it is very useful to mathematicians.
But let us apply it. You can use it to solve the following exercises.
Lemma: Let D be a unique factorization domain, and let p be an irreducible
element of D. If a,b are in D and p|ab, then p|a or p|b.
Definition: Let D be a unique factorization domain. A non-constant polynomial
f(x) = a xn + a xn-1 + · · · + a x + a in D[x]
n n-1 1 0

is called primitive if there is no irreducible element p in D such that p | a for all i.

i

LOVELY PROFESSIONAL UNIVERSITY

23
 Unit 3 : Unique Factorization Domains

Lemma: The product of two primitive polynomials is primitive. Notes

Lemma: Let Q be the quotient field of D, and let f(x) be a polynomial in Q[x]. Then
f(x) can be written in the form f(x) = (a/b)f*(x), where f*(x) is a primitive element of
D[x], a,b are in D, and a and b have no common irreducible divisors. This
expression is unique, up to units of D.
Lemma: Let D be a unique factorization domain, let Q be the quotient field of D,
and let f(x) be a primitive polynomial in D[x]. Then f(x) is irreducible in D[x] if and
only if f(x) is irreducible in Q[x].
Theorem 6: If D is a unique factorization domain, then so is the ring D[x] of
polynomials with coefficients in D.
Corollary: For any field F, the ring of polynomials
F[x , x , ... , x ]
1 2 n

in n indeterminates is a unique factorization domain.

For example, the ring Z [ 5 ] is not a unique factorization domain.

Self Assessment

1. If R is a UFD and a  R, with a  0 and being a .................., then a can be written as a product of finite number
of irreducible elements.
(a) invertible (b) non-invertible
(c) external (d) infinite
2. Any euclidean domains is a PID, it is also a ..................
(a) integral domain (b) UFD
(c) SFD (d) Ideal
3. Let R be a UFD. Then R(x) is a ..................
(a) UFD (b) SFD
(c) PID (d) Special range domain
4. In a UFD an element is prime iff it is ..................
(a) reducible (b) finite
(c) irreducible (d) infinite
5. Any two elements in a .................. have g.c.d.
(a) SFD (b) PID

(c) UFD (d) Domain

3.2 Summary

 In a PID every prime ideal is a maximal ideal.

 The definition and examples of a unique factorisation domain (UFD).
 Every PID is a UFD, but the converse is not true. Thus, Z, F’ and F[x] are UFDs, for any field F.

LOVELY PROFESSIONAL UNIVERSITY

24
 Notes  In a UFD (and hence, in a PID) an element is prime iff it is irreducible.
 Any two elements in a UFD have a g.c.d.
 If R is a UFD. then so is K[x].

3.3 Keyword

Unique Factorisation Domain: We call an integral domain R a unique

factorisation domain (UFD, in short) if every non-zero element of R which is not
a unit in R can be uniquely expressed as a product of a finite number of irreducible
elements of R.

3.4 Review Questions

1. Directly prove that F[x] is a UFD, for any field F.

(Hint: Suppose you want to factorise f(x). Then use induction on deg f(x)).
2. Give two different prime factorisations of 10 in Z.

3. Give two different factorisations of 6 as a product of irreducible elements in Z[ 5].

4. Give an example of a UFD which is not a PID.

5. If p is an irreducible element of a UFD R, then is it irreducible in every quotient ring of
R?
6. Is the quotient ring of a UFD a UFD? Why?
7. Is a subring of a UFD a UFD? Why?

Answers: Self Assessment

1. (b) 2. (b) 3. (a) 4. (c) 5. (c)

3.5 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

25
Richa Nandra, Lovely Professional University Unit 4 : Polynomial Rings

Unit 4 : Polynomial Rings Notes

CONTENTS
Objectives
Introduction
4.1 Ring of Polynomials
4.2 Some Properties of R[x]
4.3 Summary
4.4 Keywords
4.5 Review Questions
4.6 Further Readings

Objectives

After studying this unit, you will be able to:

 Identify polynomials over a given ring
 Prove and use the fact that R [x], the set of polynomials over a ring R, is a ring
 Relate certain properties of R[x] to those of R

Introduction

In the earlier units, you must have come across expressions of the form x+1, x 2+2x+1, and so on.
These are examples of polynomial. You have also dealt with polynomials in the course of Linear
Algebra. In this unit, we will discuss sets whose elements are polynomials of the type a + a, x +
0
... + a xn, where a , a ,......, a,, are elements of a ring R. You will see that this set, denoted by R [x],
n 0 1
is a ring also.
You may wonder why we are talking of polynomial rings in a block on domains and
fields. The reason for this is that we want to focus on a particular case, namely, R
[x], where R is a domain. This will turn out to be a domain also, with a lot of useful
properties. In particular, the ring of polynomials over a field satisfies a division
algorithm, which is similar to the one satisfied by Z. We will prove this property and
use it to show how many roots any polynomial over a field can have.

4.1 Ring of Polynomials

As we have said above, you may already be familiar with expressions of the type 1
+ x, 2 + 3x + 4x2, x5-1, and so on. These are examples of polynomials over the ring
Z. Do these examples suggest to you what a polynomial over any ring R is ? Let’s
hope that your definition agrees with the following one.
Definition: A polynomial over a ring R in the indeterminate x is an expression of the form
a x0 + a x1 + a x2 + ... + a xn,
0 1 2 n

LOVELY PROFESSIONAL UNIVERSITY

26
 Notes where n is a non-negative integer and a , a,, ..., a  R.
0 n

While discussing polynomials we will observe the following conventions. We will

(i) write x0 as 1, so that we will write a for a x0,
0 0

1
(ii) write x as x,
(iii) write xm instead of 1 .xm (i.e., when a = l),
m

m
(iv) omit terms of the type O.x .
Thus, the polynomial 2 + 3x2 – 1.x3 is 2x0 + 0.x1 + 3x2 + (–1)x3.
Henceforth, whenever we use the word polynomial, we will mean a polynomial in the
n
indeterminate x. We will also be using the shorter notation aixi for the polynomial
i0
a + a x+ ... + a xn.
0 1 n

Let us consider a few mox basic definitions related to a polynomial.

Definition: Let a + a, x + ... + a, xn be a polynomial over a ring R. Each of a ,a , . . ., a, is a
coefficient
n 0 l

of this polynomial. If a,  0, we call a, the leading coefficient of this polynomial.

If a = 0 = a = ... = a , we get the constant polynomial, a . Thus, every element of R is a
constant
1 2 n 0

polynomial.
In particular, the constant polynomial 0 is the zero polynomial.
It has no leading coefficient.
Now, there is a natural way of associating a non-negative integer with any non-zero polynomial.

Definition: Let a,, + a, x + . . . + a,, xn be a polynomial over a ring R, where a, 

0. Then we call the integer n the degree of this polynomial, and we write
n
 
deg aix
i
  n,of a,,  0.
 i0 
We define the degree of the zero polynomial to be – . Thus, deg 0 = –.
Let us consider some examples.
(i) 3x2 + 4x + 5 is a polynomial of degree 2, whose coefficients belong to the ring of
integers Z. Its leading coefficient is 3.
(ii) x2 + 2x4 + 6x + 8 is a polynomial of degree 4, with coefficients in Z and leading
coefficient 2. (Note that this polynomial can be rewritten as 8 + 6x + x 2 + 2x4.)
(iii) Let R be a ring and r  R, r  0. Then r is a polynomial of degree 0, with leading
coefficient r.
Before giving more examples we would like to set up some notation.
Notation: We will denote the set of all polynomials over a ring R by R[x]. (Please
note the use of the square brackets [ ]. Do not use any other kind of brackets
because R [x] and R (x) denote different sets.)
n
 i 
a x
Thus, R[x] =  1
a i  R  i  0, 1,...n, where n  0, n  Z.
i0 

We will also often denote a polynomial a + a x + . . . + a x n by f(x), p (x), q(x), etc.

0 1 n

LOVELY PROFESSIONAL UNIVERSITY

27
 Unit 4 : Polynomial Rings

Thus, an example of an element from Z [x] is f(x) = 2 x2 + 3 x + i. Notes

4

Here deg f(x) = 2, and the leading coefficient of f(x) is 2 .

Now, for ring R, we would like to see if you can define operations on the set R [x] so
that it becomes a ring. For this purpose we define the operations of addition and
multiplication of polynomials.
Definition: Let f(x) = a,, + a1x + .. + a, xn and g (x) = b + b, x + .. + b xm be two polynomials in
0 m
R[X]. Let us assume that m 2 n. Then their sum f(x) + g(x) is given by
f(x) + g(x) = (a,, + b ) + (a, + b )x + .. + (a + b,) x + b xn+l .. + b xm.
0 1 n n n+1 m

For example, consider the two polynomials p(x), q(x).in Z[x] given by

p(x) = 1 + 2x + 3x2, q(x) = 4 + 5x + 7x3

Then

p(x) + q(x) = (1+4) + (2+5)x + (3+0) x2 + 7x3 = 5 + 7x + 3x2 + 7x3.

Note that p (x) + q (x)  Z [X] and that
From the definition given above, it seems that deg (f(x)+g(x)) = max (deg f (x), deg
g (x)). But this is not always the case. For example, consider p(x) = 1 + x 2 and q (x)
= 2 + 3x – x2 in Z [X].
Then p(x) + q(x) = (1+2) + (0+3)x + (1-1)x2 = 3 a 3x.
Here deg (p(x) + q (x)) = 1 < max (deg p(x), deg q(x)).
So, what we can say is that
deg (f(x) + g(x))  max (deg f(x), deg g(x))

 f(x), g(x)  R[x].

Now let us define the product of polynomials.
Definition: If f(x) = a,, + a x + .. + a, xn and g(x) = b + b, x + .. + b xm are two polynomials in R [x],
1 0 m
we define their product f(x). g(x) by
f(x) . g(x) = c a c x +.. + c xm+n
0 1 m+n

where c = a b , + a b + .... a b  i = 0,l ,... ; m + n.

1 1 0 i-1 1 0 i

Note that a = 0 for i > n and b = 0 for i > m,

i i

As an illustration, let us multiply the following polynomials in Z[x] :

p(x) = 1 – x + 2x3, q(x) = 2 + 5x + 7x2.

Here a, = 1, a, = –1, a = 0, a = 2, b = 2, b, = 5, b = 7.
2 3 0 2

5
Thus, p(x) q(x) =  cixi , where
i=0

c = a b = 2,
0 0 0

c = a b + a b = 3,
1 10 01

c = a b + a b + a b = 2,
2 2 0 l 1 0 2

c = a b + a b + a b + a b = –3 (since b = 0).
3 30 21 12 03 3

LOVELY PROFESSIONAL UNIVERSITY

28
 Notes c =ab 0 +ab +ab +a b 3 + a b 4 = 10 (since a = 0 = b ).
4 4 3 1 2 2 1 0 4 4

c =ab +a b +ab + a b +a b + a b, = 14 (since a =0=b

5 5 0 4 1 3 2 2 3 1 4 0 5

2 3 4 5
So p(x). q(x) = 2 + 3x +2x - 3x + 10x + 14x .
Note that p(x). q(x)  Z[X], and deg (p(x) q(x)) = 5 = deg p (x) + deg q (x).
As another example, consider

p(x)  1  2x,q(x)  2  3x2  Z6[x].

Then, p(x). q(x) = 2 + 4x + 3x2 + 6x3 = 2 + 4x + 3x2 .

Here, deg (p(x). q(x)) = 2 < deg p (x) + deg q (x) (since deg p (x) = 1, deg q (x) = 2).
In the next section we will show you that
deg (f(x) g(x))  deg f(x) + deg g(x)
By now you must have got used to addition and multiplication of polynomials. We
would like to prove that for any ring R, R[x] is a ring with respect to these
operations. For this we must note that by definition, + and . are binary operations
over R [x].
Now let us prove the following theorem. It is true for any ring, commutative or not,
Theorem 1: If R is a ring, then so is R[x], where x is an indeterminate.
Proof: We need to establish the axioms R1 – R6 of Unit 14 for (R[x], + , .).
(i) Addition is Commutative: We need to show that p(x) + q(x) = q(x) + p(x) for any
p(x) , q(x)  R [x].
Let p (x) = a + a x + ... + a,xn, and
0 1

q(x) = b + b x + ... + b xm be in R[x].

0 1 m

Then, p (x) + q(x) = c + c x + ... + c xt,

0 1 1

where c = a + b and t = max (m,n).

i i i

Similarly,
q(x) + p(x) = d + d x + ... + d xs,
0 1 s

where d = b + a , s = max (n, m) = t.

i i i

Since addition is commutative in R, c, =d  i  0.

i

So we have
p(x) + q(x) = q(x) + p(x).
(ii) Addition is Associative: Again, by using the associativity of addition in R, we can
show that if p(x), q(x), s(x)  R[x], then
{p(x) + q (x)} + s(x) = p(x) + {q(x) + s(x)}.
(iii) Additive Identity: The zero polynomial is the additive identity in R [x]. This is
because,
for any p(x) = a + a, x+ ... + a xn  R[x],
0 n

0 + p(x) = (0 + a,) + (0 +a )x + ... +(0 + a )xn

1 n

= a + a x + ... +a xn
0 l n

= p(x)

LOVELY PROFESSIONAL UNIVERSITY

29
 Unit 4 : Polynomial Rings

(iv) Additive Inverse: For p (x) = a, + a x +... + a xn  R[x], consider the polynomial –p(x) =Notes
1 n
–a, –a x – ... –a xn, – a being the additive inverse of ai in R. Then
1 n i

p(x) + (–p(x)) = (a,, –a,,) + (a – a ) x + ... + (a – a )xn

1 1 1n

= 0 + 0.x + 0.x2+ ... + 0.xn

= 0.
Therefore, – p(x) is the additive inverse of p(x).
(v) Multiplication is Associative:
Let p(x) =a, +a x + ... +a xn,
1 n

q(x) = b + b x+ ... +b xm,

0 1 m

and t (x) = d +d x + ... + d xr, be in R [x]

0 1 r

Then
p(x) . q(x) = c + c x + .. . + c xs where s = m + n and
0 1 s ,

ck = a b + a b + ... + a b  k = 0,l, ...,s .

k 0 k–1 1 0 k

Therefore,
{p(x) . q(x)} t (x) = e + e x + ... +e xt,
0 1 1

where t = s + r = m+n+r and

e = c d + c d + ... + c d
k k 0 k-1 1 0 k

= (a b + ... + a b )d + (a ,b + ... + a b ) d + ... + a b d .

k 0 0 k 0 k-1 0 0 k-1 1 0 0 k

k
Similarly, we can show that the coefficient of x (for any k  0) in p(x) (q (x) t(x))
is a b d + a , (b d + b d ) + ... + a (b d + b , d + ... + b d )
k 0 0 k-1 1 0 0 1 0 k 0 k-1 1 0 k

= e , by using the properties of + and . in R.

k

Hence, {p(x).q(x)} . t(x) = p(x) . {q (x). t (x)}

(vi) Multiplication Distributes over Addition:
Let p(x) = a +a x + ... + a x ,
0 1 n n

q(x) = b + b x + ...+ b xm
0 l
m

and t(x) = d + d, x + . . . + d xr be in R[x],

0 r

The coefficient of xk in p (x). (q(x) + t (x)) is

c=a (b + d ) + a(b + d ) + (b + d ) + ... + a, (b + d ).
k k 0 0 1 1 1 1 k k

And the coefficient of xk in p (x) q (x) + p (x) t(x) is

(a b + a b + ... + a b ) + (a d +a d + ... +a d ),
k 0 k-1 1 0 k k 0 k-1 1 0 k

= a (b + d ) +a (b + d ) + ... +a (b +d)=c
k 0 0 k-1 1 0 k k k

This is true  k  0.

Hence, p (x) . (q(x) + t (x) ] = p (x) . q(x) + p (x). t.(x).

Similarly, we can prove that
{q(x)+t(x).p(x)=q(x).p(x)+t(x).p(x)
Thus, R [x] is a ring.

LOVELY PROFESSIONAL UNIVERSITY

30
 Notes Note that the definitions and theorem in this section are true for any ring. We have not
restricted
ourselves to commutative rings. But, the case that we are really interested in is
when R is a domain. In the next section we will progress towards this case.

4 .2 Some Properties of R[x]

In the previous section you must have realised the intimate relationship between
the operations on a ring R and the operations on R [x]. The next theorem reinforces
this fact.
Theorem 2: Let R be a ring.
(a) If R is commutative, so is R [x].
(b) If R has identity, so does R [x].
Proof: (a) Let p (x) = a + a, x + .. . + a,xn and
0

q(x) = b + b x + ... + b x be in R[x].

0 1 m m

Then (x) . q(x) = c + c x + .. . + c xs, where s = m + n and

0 1 s

c = a b + a b +. . .+ a b
k k 0 k-1 1 0 k

= b a + b a 4 ... + b a + b a , since both addition and multiplication are commutative

in R.
k 0 k-1 1 1 k-1 0 k

k
= coefficient of x in q (x) p(x).
Thus, for every i  0 the coefficients of x1 in p(x) q(x) and q(x) p(x) are equal
Hence, P(x) q(x) = q(x) p(x).
(b) We know that R has identity 1. We will prove that the constant polynomial 1 is the
identity of R [X]. Take any
p(x) = a + a x + ... + a xn  R[x].
0 1 n

Then 1.p(x) = c + c x + ... + c xn (since deg 1 = 0),

0 1 n

where c = a , 1 + ak .0+a
k k -1 k-2

Thus, 1, p(x) = p (x).

Similarly, p(x). 1 = p(x).
This shows that 1 is the identity of R[x]. ,
In the following exercise we ask you to check if the converse of Theorem 2 is true.
Now let us explicitly state a result which will help in showing us that R is a domain
iff R [x} is a domain, This result follows just from the definition of multiplication of
polynomials.
Theorem 3: Let R be a ring and f (x) and g (x) be two non-zero elements of R [x].
Then deg (f(x) g (x))  deg f(x) + deg g (x), with equality if R is an integral domain.
Proof: Let f (x) = a + a x + ... + a xn, a  0,
0 1 n n

and g (x) = b + b x + ... + b xm, b  0.

0 1 m m

Then deg f (x) = n, deg g (x) = m. We know that

f (x). g (x) = c +c x + ... + c × m+n
,
0 1 m+n

where c = a b + a, b + ... + a b ,.
k k 0 1 0 k

LOVELY PROFESSIONAL UNIVERSITY

31
 Unit 4 : Polynomial Rings

Since a , a , ... and b ,b . .. . are all zero, Notes

n+1 n+2 m+1 m+2
c =ab
m+n n m

Now, if R is without zero divisors, then a b  0, since a,  0

n m

and b  0. Thus, in this case,

deg (f(x) g (x)) = deg f(x) + deg g (x).
On the other hand, if R has zero divisors, it can happen that a,b, = 0.

In this case, deg (f (x) g (x)) < m+n = deg f(x) + deg g(x). Thus, our

theorem is proved.

The following result follows immediately from Theorem 3.

Theorem 4: R [x] is an integral domain <=>. R is an integral domain.
Proof: From Theorem 2, we know that R is a commutative ring with identity iff R[x]
is a commutative ring with identity. Thus, to prove this theorem we need to prove
that. R is without zero divisors iff R [x] is without zero divisors.
So let us first assume that R is without zero divisors.
Let p(x) = a + a x+ ... + a x , and q(x) = b + b x +... +b x
0 l n n 0 1 m m

be in R [x], where a,  0 and b,  0.

Then, in Theorem 3 we have seen that deg .(p (x) q (x)) = m + n  0.
Thus, p (x) q (x)  0
Thus, R [x] is without zero divisors.
Conversely, let us assume that R [x] is without zero divisors. Let a and b be non-
zero elements of R. Then they are non-zero elements of R [x] also. Therefore, ab 
0. Thus, R is without zero divisors. So, we have proved the theorem.
Now, you have seen that many properties of the ring R carry over to R’[x]. Thus, if F
is a field, we should expect F[x] to be a field also, But this is not so. F[x] can never
be a field.
This is because any polynomial of positive degree in F|x| does not have a
multiplicative inverse. Let us see why.
Let f (x)  F [x] and deg f (x) = n > 0. Suppose g (x)  F [x] such that f (x) g (x) = 1. Then
0 = deg 1 = deg (f(x) g (x)) = deg f(x) + deg g (x), since F [x] is a domain.
= n + deg g (x)  n > 0. We reach a contradiction. Thus, F [x] cannot be a field.
But there are several very interesting properties of F [x], which are similar to those
of Z, the set of integers. In the next section we shall discuss the properties of
division in F [x].

Self Assessment

1. A polynomial over a ring R in determinate X is an expression of the form .................

(a) a x0 + a x 1 + a x 2 + ...... a xn (b) a0x + a2x + a3x +...... anx
0 1 2 n 1 2 3 n

(c) a-1x + a-1x + a-1x ...... a-1x (d) a x-1 + a x-1 + a x-3 ...... a x-n
2 3 n 0 1 2 n

LOVELY PROFESSIONAL UNIVERSITY

32
 Notes 2. The degree of the zero polynomial to be ................. thus degree 0 = .................
(a) –,– (b) , 
(c) , –1 (d) –1, 
3. 3x2 + 4x + 5 is a polynomial of degree ................., whose coefficients belong to the
ring of integers Z its leading coefficients is .................
(a) 4, 5 (b)
(c) 2, 4 (d)
2 4
4. x + 2x + 6x + 8 is a polynomial of degree ................. with coefficient 2.
(a) 4 (b)
(c) 6 (d)
5. Let R be a ring and r  R, R ................. 0. Then r is polynomial degree of 0 with
leading coefficient r.
(a) = (b) 
(c)  (d) 

4.3 Summary

 The definition and examples of polynomials over a ring.

 The ring structure df R[x], where R is a ring.
 R is a commutative ring with identity iff R[x] is a commutative ring with identity.
 R is an integral domain iff R[x] is an integral domain.

4.4 Keywords

Polynomial: A polynomial over a ring R in the indeterminate x is an expression of the

form
a x0 + a x1 + a x2 + ... + a xn,
0 1 2 n

where n is a non-negative integer and a , a,, ..., a  R.

0 n

Coefficient of Polynomial: Let a + a, x + ... + a, xn be a polynomial over a ring R. Each

of a ,a , .
n 0 l

. ., a, is a coefficient of this polynomial. If a,  0, we call a, the leading coefficient of this

polynomial.

4.5 Review Questions

1. Identify the polynomials from the following expressions. Which of these are elements
of Z[x]?
2 1
(a) x6 + x5 + x4 + x2 + x + 1 (b) 
x2 x

1
(c) 3x2  2x  5 (d) 1  x
2
(e) x1/2 + 2x3/2 + 3x5/2 (f) –5

LOVELY PROFESSIONAL UNIVERSITY

33
 Unit 4 : Polynomial Rings

2. Calculate Notes
(2 + 3x2 + 4x3) + (5x + x3) in Z[x]. (b)
(a) (6  2x2 )  (1  2x  5x3 ) in Z7 [x].

(c) (1 + x) (1 + 2x + x2) in Z[x]. (d) (1  x)(1  2x  x2 ) in Z3[x]

(e) (2 + x + x2) (5x + x3) in Z[x]
3. If R is a ring such that R[x] is commutative and has identity, then
(a) is R commutative?
(b) does R have identity?

Answers: Self Assessment

1. (a) 2. (a) 3. (b) 4. (a) 5. (b)

4.6 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

34
 Sachin Kaushal, Lovely Professional University

Notes Unit 5 : Division of Algorithm

CONTENTS
Objectives
Introduction
5.1 The Division Algorithm
5.2 Summary
5.3 Keywords
5.4 Review Questions
5.5 Further Readings

Objectives

After studying this unit, you will be able to:

 Prove and use the division algorithm for F[X], where F is a field
 Discuss examples related to algorithms

Introduction

In the last unit, you have studied about polynomials rings. In this unit, we will
discuss the division of algorithm.

5.1 The Division Algorithm

We have discussed various properties of divisibility in Z. In particular, we proved

the division algorithm for integers. We will now do the same for polynomials over a
field F.
Theorem 1 (Division Algorithm): Let F be a field. Let f(x) and g(x) be two
polynomials in F[x], with g(x)  0. Then
(a) there exist two polynomials q(x) and r (x) in F [X] such that f (x) = q (x) g (x) + r (x),
where deg r(x) < deg g (x).
(b) the polynomials q (x) and r (x) are unique.
Proof: (a) If deg f (x) < deg g (x), we can choose q (x) = 0.
Then f(x) = 0. g(x) + f (x), where deg f(x) < deg g (x).
Now, let us assume that deg f(x)  deg g (x).
Let f(x) = a + a x + . . . +a xn, a,  0, and
0 1 n

g(x) = b + b x + ... + b xm, b  0, with n  m.

0 1 m m

LOVELY PROFESSIONAL UNIVERSITY

35
 Unit 5 : Division of Algorithm

We shall apply the principle of induction on deg f(x), i.e., n. Notes

If n = 0, then rn = 0, since g(x)  0. Now
f(x) = a,,, g(x) = b,, and hence
f(x) = (a,, b -l) b + 0 = q (x) g (x) + r (x), where q(x) = a b -1
and r(x) = 0.
0 0 0 0

Thus,
f(x) = q(x) g(x) + r (x), where deg r(x) < deg g(x).
So the algorithm is true when n = 0. Let us assume that the algorithm is valid for all
polynomials of degree  n – 1 and how to establish that it is true for f(x). Consider the
polynomial
f (x) = f(x) – a b -1 xn-m g(x)
1 n m
-1
= (a,, + a, x +. . .+a xn) – (a b -1 b xn-m+a b b xn-m+1 +...+ a b -1b xn)
n n m 0 n m 1 n m m

Thus, the coefficient of xn in f, (x) is zero; and hence,

deg f, (x)  n-1.
By the induction hypothesis, there exist q, (x) and r (x) in
F[x] such that f, (x) = q, (x) g(x) + r(x), where deg r(x) < deg g(x).
Substituting the value of f,(x), we get
-1
f(x)-a b g(x) = q (x) g(x) + r(x),
n m 1
-1
i.e., f(x) = {a b xn-m +q (x)} g(x) + (x)
n m 1
-1
= q(x) g(x)+r(x), where q(x) = a b xn-m +q (x)
n m 1

and deg r(x) < deg g(x).

Therefore, the algorithm is true for f(x), and hence for all polynomials in F[x].
(b) Now let us show that q(x) and r(x) are uniquely determined. If possible, let
f(x) = q (x) g(x) + r (x), where deg r (x) < deg g(x).
1 1 1

and
f(x) =q (x) g(x)+r (x), where deg r (x) < deg g(x).
2 2 2

Then
q (x) g(x)+r (x) = q (x) g(x)+r (x), so that
1 1 2 2

{q (x) – q (x)} g(x) = r (x) – r (x) ...(1)

1 2 2 1

Now if q (x)  q (x), then deg {q (x) – q (x)}  0, so that

1 2 1 2
deg [{q (x) – q (x) g(x)]  deg g(x).
1 2

On the other hand, deg {r (x) - r (x)} < deg g(x), since
2 1

deg r (x) < deg g(x) and deg r (x) < deg g(x).
2 1

But this contradicts Equation (1). Hence. Equation (1) will remain valid only if
q (x) – q (x) = 0. And then r (x) – r (x) = 0,
1 2 2 1

i.e., q (x) = q (x) and r (x) = r (x).

1 2 1 2

LOVELY PROFESSIONAL UNIVERSITY

36
 Notes Thus, we have proved the uniqueness of q(x) and r(x) in the expression f(x) = q(x) g(x)+r(x).
Here
q(x) is called the quotient and r(x) is called the remainder obtained on dividing f(x) by
g(x).
Now, what happens if we take g(x) of Theorem 1 to be a linear polynomial? We get
the remainder theorem. Before proving it let us set up some notation.
Notation: Let R be a ring and f(x)  R[x]. Let
f(x) = a + a x + ... +a xn.
0 l n

Then, for any r  R, we define

that is, f(r) is the value of f(x) obtained by substituting r for x.
Thus, if f(x) = 1 + x + x2  Z[x], then
f(2) = 1 + 2 + 4 = 7 and f(0) = 1 + 0 + 0 = 1.
Let us now prove the remainder theorem, which is a corollary to the division algorithm.
Theorem 2 (Remainder Theorem): Let F be a field. If f(x)  P[x] and b  F, then
there exists a unique polynomial q(x)  F[x] such that f(x) = (i-b) q(x)+f(b).
Proof: Let g(x) = x-b. Then, applying the division algorithm to f(x) and g(x), we can
find unique q(x) and r(x) in F[x], such that
f(x) = q(x)g(x) + r(x)
= q(x) (x – b) + r(x), where deg r(x) < deg g(x) = 1. Since deg r(x) < 1, r (x) is an
element of F, say a.
So, f(x) = (x - b)q(x) +
a, Substituting b for x,
we get f(b) = (b – b)
q(b) + a
= 0.q(b) + a= a
Thus, a = f(b).
Therefore, f(x) = (x-b) q(x)+f(b).
Note that deg f(x) = deg(x-b)+deg q(x) = l+deg q(x).
Therefore, deg q(x) = deg f(x)-1.
Let us apply the division algorithm in a few situations now.

Example: Express x4 + x3 + 5x2 – x as

(x2 + x + 1) q(x) + r(x) in Q[x].
Solution: We will apply long division of polynomials to solve this problem.
2 4 3 2

x + x + 1) x + x + 5x x

4x2  x
4x2  4x  4
 5x  4

LOVELY PROFESSIONAL UNIVERSITY

37
 Unit 5 : Division of Algorithm

Now, since the degree of the remainder -5x- 4 is less than deg .(x2+x+1), we stop the process. We Notes
get

x4 + x3 + 5x2 – x = (x2 + x + 1) (x2 + 4) – (5x + 4)

Here the quotient is x2 + 4 and the remainder is – (5x+4).
Now, let us see what happens when the remainder in the expression f = qg + r is zero.

Self Assessment

1. Let F be a field. Let f(x) and g(x) be two polynomials is f[x], with g(x)  0, then the polynomial q(x) and r(x)
an ...................
(a) unique (b) deficient
(c) finite (d) infinite
2. If deg f(x) < deg g(x) we can chosen q(x) = 0. Then f(x) = 0.g(x) + f(x) where degf(x) ...................
deg g(x).
(a) < (b) >
(c)  (d) 
4 3 2
3. x + x + 5x – x is equal to ...................
(a) (x2 + x + 1) (q(x) + r(x) is Q[x])
(b) (x + x2 + 1) (q-1(x) + r-1(x) in Q[x])
(c) (x + x2 + 1)-1 (q(x) + r(x) in Q[x])
(d) q(x)-1 + q(x)2 + (x + x2 + 1) in Q[x]
4. ................... theorem said that let F be a field, if F[x]  P[x] and b  F, then there exists a
unique polynomial q(x)  F[x] such that f(x) = (i - b) q(x) + F(b)
(a) remainder theorem (b) division algorithm
(c) contradiction theorem (d) division matrix

5.2 Summary

 The division algorithm in F[x], where F is a field, which states that if f(x), g(x)  F(x), g(x)  0, then there exist
unique q(x), r(x)  F[x] with f(x) = q(x) g(x)+r(x) and deg r(x) < deg g(x).
a F is a root of f(x)  F[x] iff (x–a) | f(x).
 A non-zero polynomial of degree n over a field F can have at the most n roots.

5.3 Keywords

Division Algorithm: Let F be a field. Let f(x) and g(x) be two polynomials in F[x], with g(x)  0.
Remainder Theorem: Let F be a field. If f(x)  P[x] and b  F, then there exists a
unique polynomial q(x)  F[x] such that f(x) = (i-b) q(x)+f(b).

LOVELY PROFESSIONAL UNIVERSITY

38
 Notes
5.4 Review Questions

1. Express f as gq + r, where deg r < degg, in each of the following cases.

(a) f = x4 + 1, g = x3 in Q[x]

(b) f = x3 + 2x2 – x + 1, g = x + 1 in Z3[x].

(c) f = x3 – 1, g = x – 1 in R[x].

p
2. You know that if p, q  Z, q  0, then can be written as the sum of an integer and a
q
fraction * with | m | < | q |. What is the analogous property for elements of F[x]?

Answers: Self Assessment

1. (a) 2. (a) 3. (a) 4. (a)

5.5 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

39
Sachin Kaushal, Lovely Professional University Unit 6 : Irreducibility and Field
Extensions

Unit 6 : Irreducibility and Field Extensions Notes

CONTENTS
Objectives
Introduction
6.1 Irreducibility in Q[x]
6.2 Field Extensions
6.2.1 Prime Fields
6.2.2 Finite Fields
6.3 Summary
6.4 Keywords
6.5 Review Questions
6.6 Further Readings

Objectives

After studying this unit, you will be able to:

 Prove and use Eisenstein’s criterion for irreducibility in Z[x] and Q[x]
 Obtain field extensions of a field F from F[x]
 Obtain the prime field of any field
 Use the fact that any finite field F has pn elements, where char F = p and dim z F = n
p

Introduction

We have discussed various kinds of integral domains, including unique factorisation

domains. Over there you saw that Z[x] and Q[x] are UFDs. Thus, the prime and
irreducible elements coincide in these rings. In this unit, we will give you a method
for obtaining the prime (or irreducible) elements of Z[x] and Q[x]. This is the
Eisenstein criterion, which can also be used for obtaining the irreducible elements
of any polynomial ring over a UFD.
After this, we will introduce you to the field extensions and subfields. We will use
irreducible polynomials for obtaining field extensions of a field F from F[x]. We will also
show you that every field is a field extension of Q or Z, for some prime p. Because of
this, we call Q and the Z s
p
prime fields. We will discuss these fields briefly.
Finally, we will look at finite fields. These fields were introduced by the young
French mathematician Evariste Galois while he was exploring number theory. We
will discuss some properties of finite fields which will show us how to classify them.
Before reading this unit ,we suggest that you go through the definitions of irreducibility.

LOVELY PROFESSIONAL UNIVERSITY

40
 Notes
6.1 Irreducibility in Q[x]

We introduced you to irreducible polynomials in F[x], where F is a field. We also

stated the Fundamental Theorem of Algebra, which said that a polynomial over C is
irreducible iff it is linear. You also learnt that if a polynomial over R is irreducible, it
must have degree 1 or degree 2. Thus, any polynomial over R of degree more than
2 is reducible. And, using the quadratic formula, we know which quadratic
polynomials over R are irreducible.
Now let us look at polynomials over Q. Again, as for any field F, a linear polynomial
over Q is irreducible. Also, by using the quadratic formula we can explicitly obtain
the roots of any quadratic polynomial over Q, and hence figure out whether it is
irreducible or not. But, can you tell whether 2x7 + 3x5 – 6x4 + 3x3 + 12 is irreducible
over Q. This criterion was discovered by the nineteenth century mathematician
Ferdinand Eisenstein. In this section we will build up the theory for proving this
useful criterion.
Let us start with a definition.
Definition: Let f(x) = a, + a x + . .. + a xn  Z[x]. We define the content of f[x] to be the g.c.d.
of the
1 n
integers a , a ,,..., a,.
0 1

We say that f(x) is primitive if the content of f(x) is 1.

For example, the content of 3x2 + 6x + 12 is the g.c.d. of 3, 6 and 12, i.e., 3. Thus,
this polynomial is not primitive. But x5 + 3x2 + 4x – 5 is primitive, since the g.c.d of
1, 0, 0, 3, 4, –5 is 1.
We will now prove that the product of primitive polynomials is a primitive
polynomial. This result is well known as Gauss’ lemma.
Theorem 1: Let f(x) and g(x) be primitive polynomials. Then so is f(x) g(x).
Proof: Let f(x) = a + a x + ... + a xn  Z[x] and
0 1 n

g(x) = b + b x + ... + b xm  Z[x], where the

0 1 m

g.c.d. of a , a , ..., a, is 1 and the g.c.d. of b , b ..., b is

0 1 0 1 m

f(x) g(x) = c + c x + ... + c xm+n

0 1 m+n

where c, = a b +ab + ... + a b .

0 k 1 k-1 k 0

To prove the result we shall assume that it is false, and then reach a contradiction. So,
suppose
that f(x) g(x) is not primitive. Then the g.c.d. of c , c ...., c is greater than 1, and hence
some
0 1 m+n
prime p must divide it. Thus, p | c  i = 0, 1, ..., m+n. Since f(x) is primitive, p does not
divide
i

some a . Let r be the least integer such that p| a . Similarly, let s be the least integer such that
i r

p| b .
s

Now consider
c =ab +ab + ... + a b + ... + a b
r+s 0 r+s I r+s-1 r s r+s 0

= a b + (a b +ab + ... + a b +a b + ... + a b)

r s 0 r+s I r+s-1 r-1 s+1 r+1 s-1 r+s 0

By our choice of r and s, p | a , p | a , ..., p | a , and p | b , p | b , ..., p | b .

0 1 r-1 0 1 s-1

Therefore, p | c – (a b +... + a b +a b + ... + a b)

r+s 0 r+s r-1 s+1 r+1 s-1 r+s 0
i.e., p | a b,.
r

 p ( a, or p | b , since p is a prime.
s
LOVELY PROFESSIONAL UNIVERSITY

41
 Unit 6 : Irreducibility and Field
Extensions

Notes
But p | a and p | b . So we reach a contradiction. Therefore, our supposition is false. That is, our
r s

theorem is true.
Let us shift our attention to polynomials over Q now.

3 1 1
Consider any polynomial over Q, say f(x) = x3 + x2 + 3x +
. If we take the l.c.m of the
2 5 3
denominators, is., of 2, 5, 1 and 3, i.e., 30 and multiply f(x) by it, what do we get? We get
30f(x) = 45x3 + 6x2 + 90x + 10  Z[x]
Using the same process, we can multiply any f(x)  Q[x] by a suitable integer d so that df(x),
 Z[X]. We will use this fact while relating irreducibility in Q[x] with irreducibility in Z[x].
Theorem 2: If f(x)  Z[x] is irreducible in Z[x], then it is irreducible in Q[x].
Proof: Let us suppose that f(x) is not irreducible over Q[x]. Then we should reach a
contradiction. So let f(x) = g(x) h(x) in Q[x], where neither g(x) nor h(x) is unit, i.e., deg
g(x) > 0, deg h(x) > 0. Since g(x)  Q[x].  m  Z such that mg(x)  Z[x]. Similarly,  n
 Z such that nh(x)  Z[x]. Then,

mnf(x) = mg(x) nh(x) ... (1)

Thus, (1) gives us
mnrf (x) = stg (x)h (x) ...(2)
1 1 1
Since g (x) and h (x) are primitive, Theorem 1 says that g (x) h (x) is primitive. Thus, the content
1 1 1 1

of the right hand side polynomial in (2) is st. But the content of the left hand side
polynomial in (2) is mnr. Thus, (2) says that mnr = st.
Hence, using the cancellation law in (2), we get f,(x) = g,(x) h (x).
1

Therefore, f(x) = rf (x) = (rg (x)) h (x) in Z[x], where neither rp (x) nor h (x) is a unit.
This
1 1 1 1 1
contradicts the fact that f(x) is irreducible in Z[x].
Thus, our supposition is false. Hence, f(x) must be irreducible in Q[x].
What this result says is that to check irreducibility of ii polynomial in Q[x], it is
enough to check it in Z[x]. And, for checking it in Z[x] we have the terrific
Eisenstein’s criterion, that we mentioned at the beginning.
Theorem 3 (Eisenstein’s Criterion): Let f(x) = a + a x + ... + a,,xn  Z[x]. Suppose that for some
0 l
prime number p;

(i) P|a,
n

(ii) p | a , p | a , ..., p | a , and

0 1 n-1

2
(iii) p |a.
0

Then f(x) is irreducible in Z[x] (and hence Q[x]).

Proof: Suppose f(x) is reducible in Z[x].
Let f(x) = g(x) h(x),
where g(x) = b + b x + ... + b,, xm, m > 0 and
0 1

h(x) = c + c x + ... + c xr,

r > 0.
0 1 r

LOVELY PROFESSIONAL UNIVERSITY

42
 Notes Then n = deg f = deg g + deg h = m + r, and

a = b c + b c ... + b c  k = 0, 1 ..., n.
k 0 k 1 k-l k 0

Now a = b c . We know that p | a . Thus, p | b c ,  p | b or p | c . Since p2 | a , p cannot divide

0 00 0 00 0 0 0

both b and c . Let us suppose that p | b and p k CJ

0 0 0

Now let us look at a,, = b, c,. Since p | a, we see that p | b and p | c . Thus, we see that for
some
m r

i, p | b . Let k be the least integer such that p | b . Note that 0 < k  m < n.
i k

Therefore, p|a .
k

Since p|a and p|b , p | b , ..., p | b , we see that p(a – (b c + .... + b c ), i. e.,
k 0 1 k–1 k 0k k–1 1

p (b c . But p | b and p | c . So we reach a contradiction.

k0 k 0

r Thus, f(x) must be irreducible in Z[x].

Let us illustrate the use of this criterion.

Example: Is 2x7 + 3x5 – 6x4 + 3x3 + 12 irreducible in Q[x]?

Solution: By looking at the coefficients we see that the prime number 3 satisfies
the conditions given in Eisenstein’s criterion. Therefore, the given polynomial is
irreducible in Q[x].

Example: Let p be a prime number. Is Q[x]/<x3 – p > a field?

Solution: You know that for any field F, if f(x) is irreducible in F[x], then <f(x)> is a
maximal ideal of F[x].
Now, by Eisenstein’s criterion, x3-p is irreducible since p satisfies the conditions
given in Theorem 3. Therefore, <x3 – p> is a maximal ideal of Q[x].
You also know that if R is a ring, and M is a maximal ideal of R, then R/M is a field.
Thus, Q[x] /<x3 – p> is a field.

Example: Let p be a prime number. Show that

f(x) = xp-1 + xp-2 + .... + x + 1 is irreducible in Z[x], f(x) is called the pth cyclotornic
polynomial.
Solution: To start with, we would like you to note that f(x) = g(x) h(x) in Z[x] iff f(x + 1) = g(x
+ 1)
h(x + 1) in Z[x]. Thus, f(x) is irreducible in Z[x] iff f(x + l) is irreducible in Z[x].

xp  1
Now, f(x) =
x 1

p
x  1  1
 f(x + 1) =
x

1
= (xp + pC xp-1 + ... + pC x + 1 – 1), (by the binomial theorem)

1 p-1

x
= xp-1 + pxp-2 + pC xp-3 + ... + pC x + p
2 p-2

LOVELY PROFESSIONAL UNIVERSITY

43
 Unit 6 : Irreducibility and Field
Extensions

Now apply Eisenstein’s criterion taking p as the prime. We find that f(x+l) is irreducible. Notes
Therefore, f(x) is irreducible.
So far we have used the fact that if f(x)  Z[x] is irreducible over Z, then it is also irreducible over
Q. Do you think we can have a similar relationship between irreducibility in Q[x] and R[xl? To

answer this, consider f(x) = x2 - 2. This is irreducible in Q[x], but f(x) = (x  2)(x  2) in R[x].
Thus, we cannot extend irreducibility over Q to irreducibility over W.
But, we can generalise the fact that irreducibility in Z[x] implies irreducibility in Q[x]. This is not
only true for Z and Q; it is true for any UFD R and its field of quotients F. Let us state this
relationship explicitly.

Theorem 4: Let R be a UFD with field of quotients F.

(i) If f(x)  R[x] is an irreducible primitive polynomial, then it is also irreducible in F[x].
(ii) (Eisenstein’s Criterion) Let f(x) = a + a x + ... + a, xn  R[x] and p  R be a prime element
0 1

such that p | a,, p2 | a0 and p | ai for 0  i < n. Then f(x) is irreducible in F[x].

The proof of this result is on the same lines as that of Theorems 2 and 3. We will
not be doing it here. But if you are interested, you should try and prove the result
yourself.
Now, we have already pointed out that if F is a field and f(x) is irreducible over F,
then F[x]/ <f(x)> is field. How is this field related to F? That is part of what we will
discuss in the next section.

6.2 Field Extensions

We shall discuss subfields and field extensions. To start with let us define these
terms. By now the definition may be quite obvious to you.
Definition: A non-empty subset S of a field F is called a subfield of F if it is a field
with respect to the operations on F. If S$F, then S is galled a proper subfield of F.
A field K is called a field extension of F if F is a subfield of K. Thus, Q is a subfield of
R and R is a field extension of Q. Similarly, C is a field extension of Q as well as of R.
Note that a non-empty subset S of a field F is a subfield of F if
(i) S is a subgroup of (F,+), and
(ii) the set of all non-zero elements of S forms a subgroup of the group of non-zero elements of F under
multiplication.
Theorem 5: A non-empty subset S of a field F is a subfield of F if and only if
(i) a  S, b  S  a – b  S, and
(ii) a  S , b  S , b  0  ab-1  S.
Now, let us look at a particular field extension of a field F. Since F[x] is an integral
domain, we can obtain its field of quotients. We denote this field by F(x). Then F is a
subfield of F(x). Thus, F(x) is a field extension of F. Its elements are expressions of
the form f,( x) where f(x), g(x)  F[x] and g(x) # 0.
g(x)
There is another way of obtaining a field extension of a field F from F[x]. We can look at
quotient rings of F[x] by its maximal ideals. You know that an ideal is maximal in F[x] iff
it is generated by an irreducible polynomial over F. So, F[x]/<f(x)> is a field iff f(x) is
irreducible over F.

LOVELY PROFESSIONAL UNIVERSITY

44
 Notes Now, given any f(x)  F[x], such that deg f(x) > 0, we will show that there is a field monomorphism
from F into F[x]/d(x)>. This will show that F[x)/<f(x)> contains an isomorphic copy of F; and
hence, we can say, that it contains F. So, let us define 0 : F  F[x]/d(x)>: (a) = a + <f(x)>.
Then  (a+b) =  (a) +  (b), and
(ab) = (a) (b).
Thus,  is a ring homomorphism.
What is Ker  ?
Ker  = {a  F] a + <f(x)> = <f(x)>}
= {a  F | a  <f(x)>}
= {a  F | f(x) | a}
= {0}, since deg f > 0 and deg a  0.
Thus,  is 1-1, and hence an inclusion.
Hence, F is embedded in F[x]/<f(x)>.
Thus, if f(x) is irreducible in F[x], then F[x]/<f(x)> is a field extension of F.
Well, we have looked at field extensions of any field F. Now let us look at certain fields, one of
which F will be an extension of.

6.2.1 Prime Fields

Let us consider any field F. Can we say anything about what its subfields look like? Yes, we
can
say something about one of its subfields. Let us prove this very startling and useful fact.
Theorem 6: Every field contains a subfield isomorphic to Q or to Z p, for some prime number
p.
Proof: Let F be a field. Define a function f : Z  F : f(n) = n.1 = 1 + 1 + .... + 1 (n times).
f is a ring homomorphism and Ker f = pZ, where p is the characteristic of F.
You know that char F = 0 or char F = p, a prime. So let us look at these two cases separately.
Case 1 (char F = 0): In this case f is one-one.  Z = f(Z). Thus, f(Z) is an integral domain
contained
in the field F. Since F is a field, it will also contain the field of quotients of f(Z). This will be
isomorphic’ to the field of quotients of Z, i.e., Q. Thus, F has a subfield which is isomorphic to Q

Case 2 (char F = p, for some prime p):

Since p is a prime number, Z/pZ is a field.
Also, by applying the Fundamental Theorem of Homomorphism to f, we get Z/pZ f(Z). Thus,
f(Z) is isomorphic to Z and is contained in F. Hence, F has a subfield isomorphic to Z .
p p

Let us Theorem 6 slightly. What it says is that:

Let F be a field.
(i) If char F = 0, then F has a subfield isomorphic to Q.
(ii) If char F = p, then P has a subfield isomorphic to Z.
Because of this property of Q arid Zp (where p is a prime number) we call these
fields prime fields.

LOVELY PROFESSIONAL UNIVERSITY

45
 Unit 6 : Irreducibility and Field
Extensions

Thus, the prime fields are Q, Z , Z , Z , etc. Notes

2 3 5

We call the subfield isomorphic to a prime field (obtained in Theorem 6), the prime
subfield of the given field.
Let us again reword Theorem 6 in terms of field extensions. What it says is that
every field is a Weld extension of a prime field.
Now, suppose a field F is an extension of a field K. Are the prime subfields of K and
F isomorphic or not? To answer this let us look at char K and char F. We want to
know if char K = char F or not. Since F is a field extension of K, the unity of F and K
is the same, namely, 1. Therefore, the least positive integer n such that n.1 = 0 is
the same for F as well as K. Thus, char K = char F. Therefore, the prime subfields of
K and F are isomorphic.
A very important fact that a field is a prime field iff it has no proper subfields.
Now let us look at certain field extensions of the fields Z .
p

6.2.2 Finite Fields

You have dealt a lot with the finite fields Z . Now we will look at field extensions of these fields.
p

You know that any finite field F has characteristic p, for some prime p. And then F is
an extension of Z. Suppose P contains q elements. Then q must be a power of p.
That is what we will prove now.
Theorem 7: Let F be a finite field having q elements and characteristic p. Then q =
pn, some positive integer n.
The proof of this result uses the concepts of a vector space and its basis.
Proof: Since char F = p, F has a prime subfield which is isomorphic to Z . We lose nothing if we
p
assume that the prime subfield is Z . We first show that F is a vector space over Z with finite
p p
dimension.
Recall that a set V is a vector space over a field K if:
(i) we can define a binary operation + on V such that (V, +) is an abelian group,

(ii) we can define a ‘scalar multiplication. : K × V  V such that  a, b  K and v, w  V, a(a + w) = a.v + a.w
(a + b).v = a.v +
b.w (ab). v = a.
(b.v)

1.v = v.
Now, we know that (F, +) is an abelian group. We also know that the multiplication
in F will satisfy all the conditions that the scalar multiplication should satisfy. Thus,
F is a vector space over 2,. Since F is a finite field, it has a finite dimension over Z .
Let dim Z F = n. Then we can
p p
find a,. .., a , a F such that
n

F = Z a + Z a + .. + Z a .
p 1 p 2 p n

We will show that F has pn elements.

Now, any element of F is of the form
b a , + b a + ... +, b a , where b,, . .., b  Zp.
1 1 2 2 n n n

Now, since o(Zp) = p, b can be any one of its p elements.

1

LOVELY PROFESSIONAL UNIVERSITY

46
 Notes Similarly, each of b , b , ...., b has p choices. And, corresponding to each of these choices
we get
2 3 n
a distinct element of F. Thus, the number of elements in F is p × p × ... × p (n times) = p n.
The utility of this result is something similar to that of Lagrange’s theorem. Using
this result we know that, for instance, no field of order 26 exists. But does a field of
order 25 exist? Does Theorem 7 answer this question? It only says that a field of
order 25 can exist. But it does not say that it does exist. The following exciting
result, the proof of which is beyond the scope of this course, gives us the required
answer. This result was obtained by the American mathematician E.H. Moore in
1893.
Theorem 8: For any prime number p and n  N, there exists a field with pn
elements. Moreover, any two finite fields having the same number of elements are
isomorphic.

Self Assessment

1. If f(x)  Z(x) is irreducible over Q[x]. Then it is .............. in Q[x].

(a) reducible (b) irreducible
(c) direct (d) finite
2. A non-empty subsets of a field F is called a .............. of F it is a field with respect to
the operation on F.
(a) subfield (b) field domain
(c) range field (d) extension
3. Every field contains a subfield is o morphic to Q or to Z for r some .............. P.
p

(a) prime (b) finite

(c) infinite (d) external
4. Let F be a finite having of elements and characteristics P, then q = .............., some
positive integer n.

(a) p-1 (b)

n
(c) xp (d)
5. For any prime number P and n  N, then exists a field with Pn elements. Move over,
and two .............. fields having the same number of elements are isomorphic.
(a) infinite (b) finite
(c) direct (d) extension

6.3 Summary

 Gauss lemma, i.e., the product of primitive polynomials is primitive.

 For any n  N, we can obtain an irreducible polynomial over Q of degree n.
 Definitions and examples of subfields and field extensions.
 Different ways of obtaining field extensions of a field F from F[x].

LOVELY PROFESSIONAL UNIVERSITY

47
 Unit 6 : Irreducibility and Field
Extensions

 Eisenstein’s irreducibility criterion for polynomials over Z and Q. This states that if Notes
f(x) = a + a, x + . . . + a xn  Z[x] and there is a prime p  Z such that
0 n

 p | a  i = 0 , 1 . ..., n – 1.
i

 p | a, and

 p | a 0,

then f(x) is irreducible over Z’ (and hence over Q).

 Every field contains a subfield isomorphic to a prime field. The prime fields are Q or Zp, for some prime p.
 The number of elements in a finite field F is pn’, where char F = p and dim z F = n.
p

 Given a prime number p’and n  N, there exists a field containing pn elements. Any two finite fields with the
same number of elements are isomorphic.

 If F is a finite field with pn elements, then xp n x is a product of pn linear polynomials over F.

6.4 Keywords
Eisenstein’s Criterion: Let f(x) = a + a x + ... + a,,xn  Z[x]. Suppose that for some prime number
0 l
p; (i) P | a , (ii) p | a , p | a ,...p|a
n 0 1 n–1

Subfield of F: A non-empty subset S of a field F is called a subfield of F if it is a

field with respect to the operations on F. If S$F, then S is galled a proper subfield of
F.

6.5 Review Questions

1. What are the contents of the following polynomials over Z?

(a) 1 + x + x2 + x3 + x4 (b) 7x4 – 7

(c) 5(2x2 – 1) (x + 2)
2. Prove that any polynomial f(x)  Z[x] can be written as dg(x), where d is the content of f(x) and g(x) is a
primitive polynomial.
3. For any n  N and prime number p, show that xn – p is irreducible over Q[x]. Note that this shows us that we can
obtain irreducible polynomials of any degree over Q[x].
4. If a + a x + ... + a, xn  Z[x] is irreducible in Q[x], can you always find a prime p that satisfied
0 1
the conditions (i), (ii) and (iii) of Theorem 3?
5. Which of the following elements of Z[x] are irreducible over Q?

(a) x2 – 12 (b) 8x3 + 6x2 – 9x + 24

(c) 5x + 1

6. Let p be a prime: integer. Let a be a non-zero non-unit square-free integer, i.e., b2 |a for any b  Z. Show that
Z[x]/<xp + a> is an integral domain.
p
7. Show that x  aZp[x] is not irreducible for any a  Z.

LOVELY PROFESSIONAL UNIVERSITY

48
Notes
Answers: Self Assessment

1. (b) 2. (a) 3. (a) 4. (b) 5. (b)

6.6 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

49
Richa Nandra, Lovely Professional University Unit 7 : Roots of a Polynomial

Unit 7 : Roots of a Polynomial Notes

CONTENTS
Objectives
Introduction
7.1 Roots of Polynomials
7.2 Summary
7.3 Keywords
7.4 Review Questions
7.5 Further Readings

Objectives

After studying this unit, you will be able to:

 Define roots of polynomials
 Discuss examples of roots of polynomial

Introduction

You have seen when we can say that an element in a ring divides another element.
Let us recall the definition in the context of F[x], where F is a field.

7.1 Roots of Polynomials

Definition: Let f(x) and g(x) be in F[x], where F is a field and g(x)  0. We say that
g(x) divides f(x)(or g(x) is a factor of f(x), or f(x) is divisible by gi(x)) if there-exists
q(x)  F[x] such that
f(x) = q(x) g(x).

We write g(x) | f(x) for ‘g(x) divides f(x)’, and g(x) | f(x) for ‘g(x) does not divide f(x)’.

Now, if f(x)  F[x] and g(x)  F[x], where g(x)  0, when g(x) | f(x)? We find that
g(x) | f(x) if r(x) = 0.
Definition: Let F be a field and f(x)  F[x]. We say that an element a  F is a root (or zero) of f(x)
if f(n) = 0.
For example, 1 is a root of x2 – 1  R[x], since 12 – 1 = 0.
1 1
Similarly, –1 is a root of f(x) = x3  x2  x   Q[x], since
2 2

1 1
f(–1) = –1 +1    0.
2 2

LOVELY PROFESSIONAL UNIVERSITY

50
 Notes Let F be a field and f (x)  F[x]. Then a  P is a root of f(x) if and only if (x–a) | f(x)).
We can generalise this criterion to define a root of multiplicity m of a polynomial in F[x].
Definition: Let F be a field and f(x)  F[x]. We say that a  F is a root of
multiplicity m (where m is a positive integer) of

f(x) if (x – a)m | f(x), but (x–a)m+1 | f(x).

For example, 3 is a root of multiplicity 2 of the polynomial (x–3) 2 (x+2)  Q[x]; and
(–2) is a root of multiplicity 1 of this polynomial.
Now, is it easy to obtain all the roots of a given polynomial? Any linear polynomial
ax+b  F[x] will have only one root, namely, -a -1b. This is because ax+b = 0 iff x =
-a-1b.
In the case of a quadratic polynomial ax 2 + bx + c  F[x], you know that its two
roots are obtained by applying the quadratic formula

b  b2  4ac
2a

For polynomials of higher degree we may be able to obtain some roots by trial and
error. For example, consider f(x) = x 5 – 2x + 1  R[x]. Then, we try out x = 1 and
find f(1) = 0. So, we find that 1 is a zero of f(x). But this method doesn’t give us all
the roots of f(x).
As we have just seen, it is not easy to find all the roots of a given polynomial. But,
we can give a definite result about the number of roots of a polynomial.
Theorem 1: Let f(x) be a non-zero polynomial of degree n over a field F:Then f(x)
has at most n roots in F.
Proof: If n = 0, then f(x) is a non-zero constant polynomial.
Thus, it has no roots, and hence, it has at most 0 ( = n) roots in F.
So, let, us assume that n  1. We will use the principle of induction on n. If deg f(x) = 1,
then
f(x) = a + a x, where a , a,  F and a,  0.
0 1 0
-1
So f(x) has only one root, namely, (–a a ).
1 0

Now assume that the theorem is true for all polynomials in F[x] of degree  n. We
will show that the number of roots of f(x)  n.
If f(x) has no root in F, then the number of roots of f(x) in F is 0 S n. So, suppose f(x)
has a root a  F.
Then f(x) = (x – a) g(x), where deg g(x) = n–1.
Hence, by the induction hypothesis g(x) has at most n–1 roots in F, say a ,....,a . Now,
1 n-1

a is a root of g(x)  g(a ) = 0  f(a ) = (a,–a) g(a ) = 0

i i i i

 a a is a root of  i = 1, ..., n – 1.
f(x)
i

Thus, each root of g(x) is a root of f(x).

Now, b  F is a root of f(x) iff f(b) = 0, i.e., iff (b – a) g(b) = 0, i.e., iff b – a = 0 or
g(b) = 0, since F is an integral domain. Thus, b is a root of f(x) iff b = a or b is a root
of g(x). So, the only roots of f(x) are a and a , ..., a . Thus, f(x) has at the most n
roots, and so, the theorem is true for n.
1 n-1

LOVELY PROFESSIONAL UNIVERSITY

51
 Unit 7 : Roots of a Polynomial

Hence, the theorem is true for all n  1. Notes

Using this result we know that, for example, x3–1  Q[x] can’t have more than 3 roots in Q.
In Theorem 1 we have not spoken about the roots being distinct. But an obvious
corollary of Theorem 1 is that
if f(x)  F[x] is of degree n, then f(x) has st most n distinct roots in F.
We will use this result to prove the following useful theorem.
Theorem 2: Let f(x) and g(x) be two non-zero polynomials of degree n over the field F. If there
exist n+l distinct elements a,,.. .,a , in F such that f(a  i = I , ..., n+l, then f(x) = g(x).
) = g(a )
n+1 i i

Proof: Consider the polynomial h(x) = f(x) = g(x)

Then deg h(x)  n, but it has n + l distinct roots a,, ..., a .
n+1

This is impossible, unless h(x) = 0, i.e., f(x) = g(x).

Example: Prove that x3  5x  Z6[x] has more roots than its degree. (Note
that Z6 is not a field.)
Solution: Since the ring is finite, it is easy for us to run through all its elements and
check which of them, are roots of

f(x)  x3  5x.

So, by substitution we find that

f(0) = 0 = f(1) = f(2) = f(3) = f(4) = f(5).

In fact, every element of Z is a zero of f(x). Thus, f(x) has 6 zeros, while deg f(x) = 3.
6

So far, we have been saying that a polynomial of degree n over F has at most n
roots in Fa. It can happen that the polynomial has no root in F. For example,
consider the polynomial x2 + 1  R[x]. You know that it can have 2 roots in R, at
the most. But as you know, this has no roots in R (it has two roots, i and –i, in C).
We can find many other examples of such polynomials in R[x]. We call such
polynomials irreducible over R. We shall discuss them in detail in the next units.
Now let us end this unit by seeing what we have covered in it.
Definition: Let F be a set on which two binary operations are defined, called
addition and multiplication, and denoted by + and · respectively. Then F is called a
field with respect to these operations if the following properties hold:
.
(i) Closure: For all a,b in F the sum a + b and the product a b are uniquely defined and belong to F.
(ii) Associative Laws: For all a,b,c in F,
a + (b + c) = (a + b) + c and a· (b· c) = (a· b)· c.
(iii) Commutative Laws: For all a,b in F,
a + b = b + a and a · b = b· a.
(iv) Distributive Laws: For all a, b, c in F,
a· (b + c) = (a· b) + (a· c) and (a + b)· c = (a· c) + (b· c).

LOVELY PROFESSIONAL UNIVERSITY

52
 Identity Elements: The set F contains an additive identity element, denoted by 0, such
Notes (v) that
for all a in F,
a + 0 = a and 0 + a = a.
The set F also contains a multiplicative identity element, denoted by 1 (and assumed to
be
different from 0) such that for all a in F,
a· 1 = a and 1· a = a.

(vi) Inverse Elements: For each a in F, the equations

a + x = 0 and x + a = 0
have a solution x in F, called an additive inverse of a, and denoted by -a. For each
nonzero
element a in F, the equations
a· x = 1 and x· a = 1

have a solution x in F, called a multiplicative inverse of a, and denoted by a -1.

, . . . , a 1,
Definition: Let F be a field. For am, am-1 a0 in F, an expression of the form
a xm + a xm-1 + · · · + a x + a
m m-1 1 0
is called a polynomial over F in the indeterminate x with coefficients a , a , . . . , a . The
set of
m m-1 0
all polynomials with coefficients in F is denoted by F[x]. If n is the largest
nonnegative integer such that a  0, then we say that the polynomial
n

f(x) = a xn + · · · + a
n 0

has degree n, written deg(f(x)) = n, and a is called the leading coefficient of f(x). If the
leading
n
coefficient is 1, then f(x) is said to be monic.
Two polynomials are equal by definition if they have the same degree and all
corresponding coefficients are equal. It is important to distinguish between the
polynomial f(x) as an element of F[x] and the corresponding polynomial function from F
into F defined by substituting elements
of F in place of x. If f(x) = a xm + · · · + a and c is an element of F, then f(c) = a c m + · · · + a . In
fact,
m 0 m 0

if F is a finite field, it is possible to have two different polynomials that define the same
polynomial
function. For example, let F be the field Z and consider the polynomials x 5 - 2x + 1 and 4x + 1.
For
5
any c in Z , by Fermat’s theorem we have c5  c (mod 5), and so
5

c5 - 2c + 1  - c + 1  4c + 1 (mod 5),
which shows that x5 - 2x + 1 and 4x + 1 are identical, as functions.
For the polynomials
f(x) = a xm + a xm-1 + · ·
m m-1

and
g(x) = b xn + b xn-1 + · · · +
n n-1

the sum of f(x) and g(x) is defined by just adding corresponding coefficients. The product
f(x)g(x)
is defined to be
a b xn+m + · · · + (a b + a b + a b )x2 + (a b + a b )x + a b .
m n 2 0 1 1 0 2
The coefficient c of xk in f(x)g(x) can be described by the formula
k

k
c = b .
k i0a i k i
LOVELY PROFESSIONAL UNIVERSITY

53
 Unit 7 : Roots of a Polynomial

This definition of the product is consistent with what we would expect to obtain using a naive Notes
approach: Expand the product using the distributive law repeatedly (this amounts to multiplying
each term be every other) and then collect similar terms.
Proposition: If f(x) and g(x) are non-zero polynomials in F[x], then f(x)g(x) is non-zero and
deg(f(x)g(x)) = deg(f(x)) + deg(g(x)).
Corollary: If f(x),g(x),h(x) are polynomials in F[x], and f(x) is not the zero polynomial, then
f(x)g(x) = f(x)h(x) implies g(x) = h(x).
Definition: Let f(x),g(x) be polynomials in F[x]. If f(x) = q(x)g(x) for some q(x) in F[x], then we say
that g(x) is a factor or divisor of f(x), and we write g(x) | f(x). The set of all polynomials divisible
by g(x) will be denoted by < g(x) >.

Lemma: For any element c in F, and any positive integer k,

(x - c) | (xk - ck).
Theorem 3: Let f(x) be a non-zero polynomial in F[x], and let c be an element of F. Then there
exists a polynomial q(x) in F[x] such that
f(x) = q(x)(x - c) + f(c).
Moreover, if f(x) = q (x)(x - c) + k, where q (x) is in F[x] and k is in F, then q (x) = q(x) and k = f(c).
1 1 1
Definition: Let f(x) = a xm + · · + a belong to F[x]. An element c in F is called a root of the
m 0
polynomial f(x) if f(c) = 0, that is, if c is a solution of the polynomial equation f(x) = 0 .
Corollary: Let f(x) be a non-zero polynomial in F[x], and let c be an element of F.
Then c is a root of f(x) if and only if x-c is a factor of f(x). That is,
f(c) = 0 if and only if (x-c) | f(x).
Corollary: A polynomial of degree n with coefficients in the field F has at most n
distinct roots in F.

Self Assessment

1. Let F be a field and f(x)  F[x] then we say that an element a  F is a root of f(x) of f(n) = ...............
(a) 1 (b) 2

(c) 0 (d) 2-1

2 2
2. 1 is a root of x – 1  R[x], since 1 – 1 = ...............
(a) 2 (b) 1
(c) 0 (d) –1

1 1

3. ............... is a root of f(x) = x3 + x2 + x   Q[x]

2 2
(a) 1 (b) 2
(c) –1 (d) –2
4. If n = 0 then f(x) is a non-zero ............... polynomial
(a) constant (b) degree
(c) range (d) power

LOVELY PROFESSIONAL UNIVERSITY

54
 Notes 5. x3 + 5x  z6[x] has ............... roots than its degree.
(a) 2 (b) 3
(c) 1 (d) more

7.2 Summary

 If f(x) and g(x) are non-zero polynomials in F[x], then f(x)g(x) is non-zero and
deg(f(x)g(x)) = deg(f(x)) + deg(g(x)).
 If f(x),g(x),h(x) are polynomials in F[x], and f(x) is not the zero polynomial, then
f(x)g(x) = f(x)h(x) implies g(x) = h(x).
 Let f(x),g(x) be polynomials in F[x]. If f(x) = q(x)g(x) for some q(x) in F[x], then we
say that g(x) is a factor or divisor of f(x), and we write g(x) | f(x). The set of all polynomials divisible by g(x)
will be denoted by < g(x) >.
 For any element c in F, and any positive integer k,

(x - c) | (xk - ck).
 Let f(x) be a non-zero polynomial in F[x], and let c be an element of F. Then there
exists a polynomial q(x) in F[x] such that
f(x) = q(x)(x - c) + f(c).
Moreover, if f(x) = q (x)(x - c) + k, where q (x) is in F[x] and k is in F, then q (x) = q(x)
and
1 1 1
k = f(c).
 Let f(x) = a xm + · · + a belong to F[x]. An element c in F is called a root of the
polynomial
m 0
f(x) if f(c) = 0, that is, if c is a solution of the polynomial equation f(x) = 0 .
 Let f(x) be a non-zero polynomial in F[x], and let c be an element of F. Then c is a
root of f(x) if and only if x-c is a factor of f(x). That is,
f(c) = 0 if and only if (x-c) | f(x).
 A polynomial of degree n with coefficients in the field F has at most n distinct roots in
F.

7.3 Keywords

Field: Let F be a set on which two binary operations are defined, called addition and
multiplication, and denoted by + and · respectively. Then F is called a field with respect
to these operations.
Identity Elements: The set F contains an additive identity element, denoted by 0,
such that for all a in F,
a + 0 = a and 0 + a = a.
Inverse Elements: For each a in F, the equations
a + x = 0 and x + a = 0
have a solution x in F, called an additive inverse of a, and denoted by -a. For each
non-zero element a in F, the equations
a· x = 1 and x· a = 1

LOVELY PROFESSIONAL UNIVERSITY

55
 Unit 7 : Roots of a Polynomial

Notes
7.4 Review Questions
1. Let F be a field and f(x)  F[x] with deg f(x)  1. Let a  F. Show that f(x) is divisible by x – a iff f(a) = 0.
2. Find the roots of the following polynomials, along with their multiplicity.

1 5
(a) f(x)  x2  x  3 Q[x] (b) f(x)  x2  x  1  Z3[x]
2 2
(c) f(x)  x4  2x3  2x  1 Z5[x]

3. Let F be a field and a  F. Define a function  : F[x]  F : f(f(x)) = f(x)

This function is the evaluation
at a. Show that
(a) f is an onto ring homomorphism.

(b) f (b) = b  b  F.
(c) Ker f = <x – a>
So, what does the Fundamental Theorem of Homomorphism say in this case?

4. Let p be a prime number. Consider xp1  1 Zp[x]. Use the fact that Zp is a group of order p to show that every
non-zero element of Zp is a root of xp-1 – 1. Thus, show that xp-1 – 1 = (x  1)(x  2)...(x  p  1).

5. The polynomial x4 + 4 can be factored into linear factors in Z [x]. Find this factorisation.
5

Answers: Self Assessment

1. (c) 2. (c) 3. (c) 4. (a) 5. (d)

7.5 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

56
 Richa Nandra, Lovely Professional University

Notes Unit 8 : Splitting Fields, Existence and

Uniqueness

CONTENTS
Objectives
Introduction
8.1 Extension Field
8.2 Summary
8.3 Keywords
8.4 Review Questions
8.5 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss splitting field
 Describe extension field and theorem related to extension

Introduction

Beginning with a field K, and a polynomial f(x)  K, we need to construct the

smallest possible extension field F of K that contains all of the roots of f(x). This will
be called a splitting field for f(x) over K. The word “the” is justified by proving that
any two splitting fields are isomorphic.
Let F be an extension field of K and let u  F. If there exists a non-zero polynomial
f(x)  K[x] such that f(u) = 0, then u is said to be algebraic over K. If not, then u is
said to be transcendental over K.

8.1 Extension Field

If F is an extension field of K, and u  F is algebraic over K, then there exists a

unique monic irreducible polynomial p(x)  K[x] such that p(u) = 0. It is the monic
polynomial of minimal degree that has u as a root, and if f(x) is any polynomial in
K[x] with f(u) = 0, then p(x) | f(x).

Alternate proof: The proof in the text uses some elementary ring theory. Then
decided to include a proof that depends only on basic facts about polynomials.
Assume that u  F is algebraic over K, and let I be the set of all polynomials f(x) 
K[x] such that f(u) = 0. The division algorithm for polynomials can be used to show
that if p(x) is a non-zero monic polynomial in I of minimal degree, then p(x) is a
generator for I, and thus p(x) | f(x) whenever f(u) = 0.
Furthermore, p(x) must be an irreducible polynomial, since if p(x) = g(x)h(x) for
g(x); h(x)  K[x], then g(u)h(u) = p(u) = 0, and so either g(u) = 0 or h(u) = 0 since
F is a field. From the choice of p(x) as a polynomial of minimal degree that has u as
a root, we see that either g(x) or h(x) has the same degree as p(x), and so p(x)
must be irreducible.

LOVELY PROFESSIONAL UNIVERSITY

57
 Unit 8 : Splitting Fields, Existence and
Uniqueness

In the above proof, the monic polynomial p(x) of minimal degree in K[x] such that p(u) = 0 is Notes
called the minimal polynomial of u over K, and its degree is called the degree of u over K.
Let F be an extension field of K, and let u , u , ..., u  F. The smallest subfield of F
that 1 2
contains
n 1 2

1 2 n
K and u , u ,..., u will be denoted by K(u , u ,..., u ). It is called the extension field of K generated
n
by u1, u ,...., u . If F = K(u) for a single element u  F, then F is said to be a simple extension of K.
2 n

Let F be an extension field of K, and let u  F. Since K(u) is a field, it must contain
all elements of the form
2 m 2 n
a 0 + a1u + a2u + ... + amu , b0 + b1u + b2u + ... + bn u
where a , b  K for i = 1,..., m and j = 1,... n. In fact, this set describes K(u), and if u is transcendental
i j

over K, this description cannot be simplified. On the other hand, if u is algebraic

over K, then the denominator in the above expression is unnecessary, and the
degree of the numerator can be assumed to be less than the degree of the minimal
polynomial of u over K.
If F is an extension field of K, then the multiplication of F defines a scalar
multiplication, considering the elements of K as scalars and the elements of F as
vectors. This gives F the structure of a vector space over K, and allows us to make
use of the concept of the dimension of a vector space. The next result describes the
structure of the extension field obtained by adjoining an algebraic element.
Definition: Let F be an extension field of K and let u  F be an element algebraic over K.
(a) K(u) K[x]= hp(x)i, where p(x) is the minimal polynomial of u over K.
(b) If the minimal polynomial of u over K has degree n, then K(u) is an n-dimensional vector space over K.
Alternate proof: The standard proof uses the ring homomorphism  : K[x]  F
defined by evaluation at u. Then the image of  is K(u), and the kernel is the ideal
of K[x] generated by the minimal polynomial p(x) of u over K. Since p(x) is
irreducible, ker() is a prime ideal, and so K[x] = ker() is a field because every
nonzero prime ideal of a principal ideal domain is maximal. Thus K(u) is a field since
K(u) 245= K[x]= ker().
The usual proof involves some ring theory, but the actual ideas of the proof are much
simpler. To give an elementary proof, define  : K[x]= {p(x)}  K(u) by ([f(x)]) = f(u),
for all congruence classes [f(x)] of polynomials (modulo p(x)). This mapping makes sense
because K(u) contains u, together with all of the elements of K, and so it must contain
any expression of the form a +a u+
0 1
... + a um, where a  K, for each subscript i. The function  is well-defined, since it is also
m i

independent of the choice of a representative of [f(x)]. In fact, if g(x)  K[x] and f(x)
is equivalent to g(x), then f(x) – g(x) = q(x)p(x) for some q(x)  K[x], and so f(u) –
g(u) = q(u)p(u) = 0, showing that ([f(x)]) = ([g(x)]).
Since the function  simply substitutes u into the polynomial f(x), and it is not
difficult to show that it preserves addition and multiplication. It follows from the
definition of p(x) that  is one-to-one. Suppose that f(x) represents a nonzero
congruence class in K[x]= {p(x)}. Then
p(x) | f(x), and so f(x) is relatively prime to p(x) since it is irreducible. Therefore, there exist
polynomials a(x) and b(x) in K[x] such that a(x)f(x) + b(x)p(x) = 1. It follows that [a(x)]
[f(x)] = [1] for the corresponding equivalence classes, and this shows that K[x] /{p(x)} is
a field. Thus the image E of  in F must be subfield of F. On the one hand, E contains u
and K, and on the other hand, we have already shown that E must contain any
expression of the form a + a u + ... + a um,
0 1 m
where a  K. It follows that E = K(u), and we have the desired isomorphism.
i

(b) It follows from the description of K(u) in part (a) that if p(x) has degree n, then the set B = {1, u, u 2,..., un-1} is a
basis for K(u) over K.

LOVELY PROFESSIONAL UNIVERSITY

58
 Let F be an extension field of K. The dimension of F as a vector space over K is called the
Notes degree
of F over K, denoted by [F : K]. If the dimension of F over K is finite, then F is said to be a finite
extension of K. Let F be an extension field of K and let u  F. The following conditions are
equivalent: (1) u is algebraic over K; (2) K(u) is a finite extension of K; (3) u belongs to a finite
extension of K.
Never underestimate the power of counting: the next result is crucial. If we have a tower of
extensions K  E  F, where E is finite over K and F is finite over E, then F is finite over K, and
[F : K] = [F : E][E : K]. This has a useful corollary, which states that the degree of any element
of
F is a divisor of [F : K].

Let K be a field and let f(x) = a0 + a1x + ... + anxn be a polynomial in K[x] of degree n > 0. An
extension field F of K is called a splitting field for f(x) over K if there exist elements r , r ,..., r
1 2
 F such that
(i) f(x) = a (x – r )(x – r ) ... (x – rn) and
n 1
(ii) F = K(r , r ..., r ).
1 2, n

In this situation we usually say that f(x) splits over the field F. The elements r , r ,..., r are
roots
1 2 n

of f(x), and so F is obtained by adjoining to K a complete set of roots of f(x). An

induction argument (on the degree of f(x)) can be given to show that splitting fields
always exist. Theorem states that if f(x)  K[x] is a polynomial of degree n > 0,
then there exists a splitting field F for f(x) over K, with [F : K]  n!.
The uniqueness of splitting fields follows from two lemmas. Let  : K  L be an
isomorphism of fields. Let F be an extension field of K such that F = K(u) for an
algebraic element u  F. Let p(x) be the minimal polynomial of u over K. If v is any
root of the image q(x) of p(x) under , and E = L(v), then there is a unique way to
extend  to an isomorphism : F  E such that (u) = v and (a) = (a) for all a  K.
The required isomorphism  : K(u)  L(v) must have the form
(a + a u + ... + a un-1) =
0 1 n-1 0

The second lemma is stated as follows. Let F be a splitting field for the polynomial f(x)  K[x]. If
 : K  L is a field isomorphism that maps f(x) to g(x)  L[x] and E is a splitting field for
g(x) over L, then there exists an isomorphism  : F  E such that (a) = (a) for all a  K. The proof uses
induction on the degree of f(x), together with the previous lemma.
Theorem states that the splitting field over the field K of a polynomial f(x)  K[x] is
unique up to isomorphism. Among other things, this result has important
consequences for finite fields.

Self Assessment

1. If F is an extension field k and u  F is algebraic over K, then their exists a ...............

(a) different (b) finite
(c) infinite (d) unique

2. The monic polynomial P(x) of minimal degree in K[x] such that P(u) = 0 is called is
............... of r over K and its degree is called the degree of u over K.
(a) maximal polynomial (b) minimal polynomial
(c) finite polynomial (d) infinite polynomial
3. The dimension of F as a vector space K is called the ............... of F over K, denoted by [f :
k]

(a) range (b) domain

(c) degree (d) field

LOVELY PROFESSIONAL UNIVERSITY

59
 Unit 8 : Splitting Fields, Existence and
Uniqueness

4. The splitting field over the field K of a polynomial f(x)  K[x] is unique up to ............... Notes
(a) homomorphism (b) isomorphism
(c) automorphism (d) finite extension

8.2 Summary

If F is an extension field of K, and u  F is algebraic over K, then there exists a

unique monic irreducible polynomial p(x)  K[x] such that p(u) = 0. It is the monic
polynomial of minimal degree that has u as a root, and if f(x) is any polynomial in
K[x] with f(u) = 0, then p(x) | f(x).

Alternate Proof: The proof in the text uses some elementary ring theory. I’ve
decided to include a proof that depends only on basic facts about polynomials.
Let F be an extension field of K, and let u , u , ..., u  F. The smallest subfield of F
that contains
1 2 n
K and u , u ,..., u will be denoted by K(u , u ,..., u ). It is called the extension field of K generated
1 2 n 1 2 n
by u , u ,...., u . If F = K(u) for a single element u  F, then F is said to be a simple extension of K.
1 2 n

Let F be an extension field of K, and let u  F. Since K(u) is a field, it must contain
all elements of the form

a 0+ a1u + a2u2 + ... + amum , b0 + b1u + b2u2 + ... + bn un

where a , b  K for i = 1,..., m and j = 1,... n. In fact, this set describes K(u), and if u is transcendental
i j

over K, this description cannot be simplified. On the other hand, if u is algebraic

over K, then the denominator in the above expression is unnecessary, and the
degree of the numerator can be assumed to be less than the degree of the minimal
polynomial of u over K.
If F is an extension field of K, then the multiplication of F defines a scalar
multiplication, considering the elements of K as scalars and the elements of F as
vectors. This gives F the structure of a vector space over K, and allows us to make
use of the concept of the dimension of a vector space. The next result describes the
structure of the extension field obtained by adjoining an algebraic element.
The uniqueness of splitting fields follows from two lemmas. Let  : K  L be an
isomorphism of fields. Let F be an extension field of K such that F = K(u) for an
algebraic element u  F. Let p(x) be the minimal polynomial of u over K. If v is any
root of the image q(x) of p(x) under , and E = L(v), then there is a unique way to
extend  to an isomorphism : F  E such that (u) = v and (a) = (a) for all a  K.
The required isomorphism  : K(u)  L(v) must have the form
(a + a u + ... + a un-1) = (a ) + (a )v + ... + 
0 1 n-1 0 1

The second lemma is stated as follows. Let F be a splitting field for the polynomial
f(x)  K[x]. If  : K  L is a field isomorphism that maps f(x) to g(x)  L[x] and E is
a splitting field for g(x) over L, then there exists an isomorphism  : F  E such that
(a) = (a) for all a  K. The proof uses induction on the degree of f(x), together
with the previous lemma.
The splitting field over the field K of a polynomial f(x)  K[x] is unique up to isomorphism.

8.3 Keywords

Splitting Field: Beginning with a field K, and a polynomial f(x)  K, we need to

construct the smallest possible extension field F of K that contains all of the roots of
f(x). This will be called a splitting field for f(x) over K.

LOVELY PROFESSIONAL UNIVERSITY

60
 Notes Extension Field: Let F be an extension field of K and let u  F. If there exists a nonzero
polynomial
f(x)  K[x] such that f(u) = 0, then u is said to be algebraic over K. If not, then u is
said to be transcendental over K.
In the above proof, the monic polynomial p(x) of minimal degree in K[x] such that
p(u) = 0 is called the minimal polynomial of u over K, and its degree is called the
degree of u over K.

8.4 Review Questions

1. Find the splitting field over Q for the polynomial x4 + 4.

2. Let p be a prime number. Find the splitting fields for xp – 1 over Q and over R.
3. Find the splitting field for x3 + x + 1 over Z .
2

4. Find the degree of the splitting field over Z for the polynomial (x 3 + x + 1)(x2 + x +
1).
2

5. Let F be an extension field of K. Show that the set of all elements of F that are
algebraic over K is a subfield of F.
6. Let F be a field generated over the field K by u and v of relatively prime degrees m
and n, respectively, over K. Prove that [F : K] = mn.
7. Let F  E  K be extension fields. Show that if F is algebraic over E and E is algebraic
over K, then F is algebraic over K.
8. Let F  K be an extension field, with u  F. Show that if [K(u) : K] is an odd number,
then K(u2) = K(u).
9. Find the degree [F : Q], where F is the splitting field of the polynomial x 3 – 11 over
the field Q of rational numbers.
10. Determine the splitting field over Q for x4 + 2.
11. Determine the splitting field over Q for x4 + x2 + 1.
12. Factor x6 – 1 over Z ; factor x5 – 1 over Z .
7 11

Answers: Self Assessment

1. (d) 2. (b) 3. (c)4. (b)

8.5 Further Readings

Books
Dan Saracino: Abstract Algebra; A First Course.
Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links
www.jmilne.org/math/CourseNotes/
www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

61
Sachin Kaushal, Lovely Professional University Unit 9 : Separable Extensions

Unit 9 : Separable Extensions Notes

CONTENTS
Objectives
Introduction
9.1 Separability

9.2 Summary

9.3 Keywords

9.4 Review Questions

9.5 Further Readings

Objectives

After studying this unit, you will be able to:

 Define separability
 Discuss examples related to separable extension

Introduction

In the last unit, you have studied about the splitting field and extension field. This
unit will provide you information related to separable extension.

9.1 Separability

Separability of a finite field extension L/K can be described in several different

ways. The original definition is that every element of L is separable over K (that is,
has a separable minimal polynomial in K[X]). We will give here three descriptions of
separability for a finite extension and use each of them to prove two theorems
about separable extensions.
Theorem 1: Let L/K be a finite extension. Then L/K is separable if and only if the trace function
Tr : L  K is not identically 0.
L/K

The trace function is discussed in Appendix A.

Theorem 2: Let L/K be a finite extension. Then L/K is separable if and only if the ring K K L

has no non-zero nilpotent elements. When L/K is separable, the ring K K L is isomorphic to
[L:K]
K .

Example: Consider the extension Q( 2 )=Q. Since Q( 2 ) Q[X]/(X2 – 2), tensoring with
Q gives Q Q Q  2  ; Q[X]/(X 2
 2)  Q[X]/((X  2)(X  2) Q  Q,

which is a product of 2 copies of Q (associated to the 2 roots of X 2) and has no nilpotent
2

elements besides 0.

LOVELY PROFESSIONAL UNIVERSITY

62
 Notes
Example: Consider the extension F ( u )/F (u). Since F ( u ) F [X]/(X2 – u),
2 2 2 2

F2 (u)  F2 (u) F2 ( u) F2 ( u) F2 (u)[X]/(X2 – u) = F2 (u)[X]/(X – u)2 ,

which has the nonzero nilpotent element X – u .

Theorem 3: Let L/K be a finite extension. Then L is separable over K if and only if
any derivation of K has a unique extension to a derivation of L.
For above two proofs, the reader should be comfortable with the fact that injectivity
and surjectivity of a linear map of vector spaces can be detected after a base
extension: a linear map is injective or surjective if and only if its base extension to a
larger field is injective or surjective.

Each of the three theorems above will be proved and then lead in its own way to
proofs of the following two theorems.
Theorem 4: If L = K(a ,....., a ) and each a is separable over K then every element of L is
separable
1 r i

over K (so L/K is separable).

Theorem 5: Let L/K be a finite extension and F be an intermediate field. If L/F and
F/K are separable then L/K is separable.
We will use our new viewpoints to define separability for arbitrary (possibly non-
algebraic) field extensions.
We want to show L/K is separable if and only if Tr : L  K is not identically 0. The trace map
L/K

is either identically 0 or it is onto, since it is K-linear with target K, so another way

of putting Theorem 1 is that we want to show L/K is separable if and only if the
trace from L to K is onto.
Proof: We might as well take K to have positive characteristic p, since in characteristic 0
all
finite field extensions are separable and the trace is not identically 0 : TrL (1) = [L : K]  0
in
/K
characteristic 0.
If L/K is separable, by the primitive element theorem we can write L = K() where  is
separable over K. To show the trace is surjective for finite separable extensions, it
suffices to prove surjectivity of the trace map on K()/K when K is any base field and  is
separable over K.
If L/K is inseparable, then there must be some a  L which is inseparable over K. Since
Tr = Tr o Tr , it success to prove the trace map on K()=K vanishes when  is
inseparable
L/K K()/K L/K()
over K.
For both cases of the field extension K()/K ( separable or inseparable over K), let  have
minimal polynomial (X) in K[X]. Write (X) =  (Xpm) where m is as large as
possible, so  (X) is separable. Thus (X) is separable if and only if m = 0.

Let n = deg  = pmd, with d = deg  . In K[X],

 (X)=(X–)...(X– )
1 d

m
for some  ’s, which are all distinct since  (X) is separable. Write   gp , so the  ’s are distinct.
i i
i i
Then
m m m m m

(X)  (Xp )  (Xp  1 ) ... (Xp d )  (X  1 )p ...(X  d )p .

LOVELY PROFESSIONAL UNIVERSITY

63
 Unit 9 : Separable Extensions

Notes
Consider now the extension of scalars up to K of the trace map Tr : K(a)  K:

K(a)/K

 Tr :
Tr  id K k(a)/K K K K(a)  K KK K.
Tr is the trace map on K K K() as a K -vector space.

Since tensoring with a field extension preserves injectivity and surjectivity of a linear map,

Tr is onto  Tr is onto; Tr =K  Tr = 0
K(a)/K K(a)

Since K() K[X] /((X)) as K-algebras, K() K[X]/((X)) as K -algebras, and thus is isomorphic
m

to the direct product of the rings K [X]/ (Xp – i ). The trace is the sum of the traces to K on each
m m

K[X]/(Xp – i ). Let’s look at the trace from K[X]/(Xp – i ). to K .

In K [X], Xpm –  =(X– )pm . Then K[X]/(Xpm –  )= K[Y]/(Ypm ), where Y = X   . If m = 0, then

i i i i
m m

K [Y ] / (Yp ) = K , so the trace to K is the identity. If m > 0, any element of K [Y ]/ (Yp ) is

the sum of a constant plus a multiple of Y , which is a constant plus a nilpotent element
(since Y mod Ypr is nilpotent). Any constant in K [Y ]/ (Ypm ) has trace 0 since pm = 0 in K
(because
m

m > 0). A nilpotent element has trace 0. Thus the trace to K of any element of K [Y ] / (Yp ) is 0.

To summarize, when  is separable over K (i.e., m = 0), the trace map from K() to K is onto since
it is onto after extending scalars to K . When a is inseparable over K (i.e., m > 0),
the trace map is identically 0 since it vanishes after extending scalars.
Theorem 1 implies Theorem 4.
Proof. Set L =K,L = K( ) = L ( ), and more generally L = K( ,.... ) = L ( ) for i  1. So we
0 1 1 0 1 i 1 i i-1 i

have the tower of field extensions

K = L L L ... L L =L.
0 1 2 r-1 r

By transitivity of the trace,

Tr = Tr o Tr o o Tr
L/K L1/L0 L2/L1 Lr/Lr–1

Since  is separable over K and the minimal polynomial of  over L divides its minimal
i i i-1

polynomial over K,  is separable over L . Therefore Tr ( )/L :L L is onto from the

i i-1 Li-1 ii-1 i i-1
proof of Theorem 1, so the composite map TrL/K : L  K is onto. Therefore L/K is
separable by Theorem 1.
Corollary: Theorem 1 implies Theorem 5.
Proof: By Theorem 1 and the hypothesis of Theorem 5, both Tr and Tr are onto. Therefore,
L/F F/K
their composite Tr is onto, so L/K is separable by Theorem 1.
L/K

Proof: We will begin with the case of a simple extension L = K( ). Let (X) be the
minimal polynomial of  over K, so L K[X]/((X)) as K-algebras and

K K L  K[X]/(X)

as K -algebras. This ring was considered in the proof of Theorem 1, where we saw its structure
is different when (X) is separable or inseparable. If (X) is separable in K[X], then K[X]/( K (X))
LOVELY PROFESSIONAL UNIVERSITY

64
 Notes
is a product of copies of the field K , so it has no non-zero nilpotent elements. If (X) is inseparable,
m

then K [X]/((X)) is a product of copies of rings K [Y ]/( Yp ) with m > 0, which all
have nonzero nilpotents.

Now we consider the structure of K  L when L/K is any finite extension.

K

First assume L/K is separable. By the primitive element theorem, we can write L =
[L:K]
K() and  is separable over K. By the first paragraph of the proof, K K L K
since (X) has distinct
linear factors in K .
If L/K is inseparable, then some a  L is inseparable over K. Tensoring the inclusion
map K()  L up to K , we have an inclusion

K  K() K  L.
K K

The ring K  K() has a non-zero nilpotent element by the first paragraph of the proof, so
K
K  L does as well.
K

Corollary: The proof of Theorem 2 implies Theorem 4.

Proof: Make a tower of intermediate extensions in L/K as in (2.2). Note K is an

algebraic closure of every field L in the tower. Since
i

KKL (KKL1)L1L

and L = K( ) with  separable over K, the proof of Theorem 2 implies

1 1 1

K K L 1 K [L1 :K]

as K -algebras. Therefore

K K L K [L1 :K]
L1 L (K L 1 L) [L

Since L = L ( ,...  ) with each  separable over L , we can run through the same
computation
1 2 r i 1

for K L2 L as we did for K K L, and we get K L1 L (K L2 L)[L2 :L1 ]
, so

K K L (K L L) 2
[L :L ][L :K]
2 1 1 = (K  L2 L)
[L :K]
2 .
Repeating this enough, in the end we get

K K L (K  Lr L)[Lr :K]

Corollary: The proof of Theorem 2 implies Theorem 5.

Proof: The field K is an algebraic closure of F and L. Using Theorem 2,

KL (KF) L
K KF

K [F:K] L since F/K is separable

F

( K  L)[F:K]
F
LOVELY PROFESSIONAL UNIVERSITY

65
 Unit 9 : Separable Extensions

Notes
K [L:F][F:K] since L=F is separable
K [L:K]

Thus L/K is separable by Theorem 1.2.

Theorem 6: Let L/K be an extension of fields, and   L be algebraic over K. Then
is separable over K if and only if any derivation on K has a unique extension to a
derivation on K().
Proof: When   L is separable over K, Corollary B.10 shows any derivation on K
extends uniquely to a derivation on K().
Now suppose   L is inseparable over K. Then ’(X) = 0, where (X) is the
minimal polynomial of  over K. In particular ’() = 0. We are going to use this
vanishing of ’() to construct a nonzero derivation on K() which extends the zero
derivation on K.
Then the zero derivation on K has two lifts to K(): the zero derivation on K() and
this other derivation we will construct.
Define Z : K()  K() by Z(f()) = f’(), where f(X)  K[X]. Is this well-defined?
Well, if f () = f (), then f (X)  f (X) mod (X), say
1 2 1 2

f (X) = f (X) + (X)k(X).

1 2

Differentiating both sides with respect to X,

f’ (X) = f’ (X) + (X)k’(X) + ’(X)k(X):
1 2

Evaluating both sides at yields f’ () = f’ () since ’() = 0. So Z : K()  K() is well-defined.
1 2

It is left to the reader to check Z is a derivation on K(). This derivation kills K, but
Z() = 1, so Z extends the zero derivation on K while not being the zero derivation
itself.
The reader can check more generally that when  is inseparable over K and  
K() is arbitrary the map f()  f’() is a derivation on K() that extends the zero
derivation on K and sends  to . So there are many extensions of the zero
derivation on K to K(): one for each element of K().
We need a lemma to put inseparable extensions into a convenient form for our
derivation constructions later.
Lemma: Let L/K be a finite inseparable field extension. Then there is an  L and
intermediate field F such that L = F() and  is inseparable over F.
Proof: Inseparable field extensions only occur in positive characteristic. Let p be
the characteristic of K. Necessarily [L : K] > 1. Since L/K is inseparable, there is
some   L that is inseparable over K.
Write L = K( ,....  ). We will show by contradiction that some  has to be inseparable over K.
1 r i

Assume every  is separable over K. Then we can treat L/K as a succession of simple field
i
extensions as in (2.2), where L = L ( ) with  separable over L . By Theorem, any derivation
i-1 i i-1 i i

on L extends to i-1 a derivation on L , so any derivation on K extends to a derivation on L. Moreover, this extended derivation on L is i unique. To show that, consider two derivations D and D’ on L that are equal on K. Since L = K( ) and  is separable over K, the proof of Corollary

1 1 1
B.10 tells us that D and D’ both send L to L and are equal on L . Now using L in place of K, D and
1 1 1 1
D’ being equal on L implies they are equal on L since L = L ( ) and  is separable over L . We
1 2 2 1 2 2 1
can keep going like this until we get D = D’ on L = L. As a special case of this uniqueness, the only
r
derivation on L which vanishes on K is the zero derivation on L.

LOVELY PROFESSIONAL UNIVERSITY

66
 Notes Now replace K as base field with K(), over which the  ’s are of course still separable. Then any
i
derivation on K() extends uniquely to a derivation on L. But in the proof of Theorem we
saw
there is a non-zero derivation Z on K() that vanishes on K, and an extension of that to a
derivation on L is non-zero on L and is zero on K. We have a contradiction of the
uniqueness of
extensions, so in any set of field generators { ,....,  }, some  must be inseparable in K.
1 r i

Choose a generating set { ,...., } with as few inseparable r r

elements 1
as r-1

possible. At least one 

1 r i
is inseparable over K and we may assume that  is one of them. Set  =  and F = K( ,.... )

(so F = K if r = 1). Then L = F(). We will show by contradiction that  must be

inseparable over F, which is the point of the lemma.
Suppose  is separable over F. Then  is separable over the larger field F(p) since
its minimal polynomial over F(p) divides its minimal polynomial over F. Since  is a
root of Xp – p
 F(p)[X], its (separable) minimal polynomial in F(p)[X] is a factor of this, so that
polynomial
must be X – . Therefore,   F(p). Taking pk-th powers for any k  0, apk F(apk+1 ), so

F(ap )  F(ap
k k+1
).

F(apk
The reverse inclusion is obvious, so ) = F(ap k+1
) for all k  0. Therefore,

L = F() = F(pk ) = K(1 ,..., r-1 , rpk )

pk
for any k  0. We can pick k so that is separable over K (why?). Then t
pk
{1 ,...,r1 ,ar } has with one less inseparable element among the field
contradicts the choice of generators to have as few members in the list as possible
that are inseparable over K, so  has to be inseparable over F.
Proof: Assume L/K is separable, so by the primitive element theorem L = K()
where  is separable over K. Any derivation on K can be extended (using
Theorem) uniquely to a derivation on L.
If L/K is inseparable, then Lemma lets us write L = F() with  inseparable over F,
and F  K. The, by a construction used in the proof of Theorem, f()  f’() with
f(X)  F[X] is a nonzero derivation on L which is zero on F, and thus also zero on the
smaller field K. This shows the zero derivation on K has a non-zero extension (and
i i-1 i i
thus two extensions) to a derivation on L.
i-1 i

Corollary: The proof of Theorem 3 implies Theorem 4.

Proof: Again we consider the tower of field extensions (2.2). Since L = L ( ) and  is separable
over L , the proof shows any derivation on L extends uniquely to a derivation on L . Therefore,
i-1
any derivation on K = L can be extended step-by-step through the tower (2.2) to a derivation on
0
Lr = L. By the argument in the proof of Lemma, this derivation on L is unique.
Lemma: Let L/K be a finite extension and F be an intermediate extension such that
F/K is separable. Then any derivation F  L which sends K to K has values in F.
Proof: Pick   F, so  is separable over K. Now use Corollary B.10 to see the
derivation F  L sends  to an element of K()  F.
Corollary: Theorem 3 implies Theorem 5.
Proof: To prove L/K is separable, we want to show any derivation on K has a unique
extension to a derivation on L. Since F/K is separable, a derivation on K extends to a
derivation on F. Since L/F is separable, a derivation on F extends to a derivation on
L.

LOVELY PROFESSIONAL UNIVERSITY

67
 Unit 9 : Separable Extensions

For uniqueness, let D and D be derivations on L which extend the same derivation on K. Since Notes
1 2
D (K)  K and D (K)  K, we have D (F)  F and D (F)  F by Lemma. Then D = D on F since
1 2 1 2 1 2
F/K is separable, and D = D on L since L/F is separable.
1 2

When L/K is an algebraic extension of possibly infinite degree, here is the way
separability is defined.
Definition: An algebraic extension L/K is called separable if every finite
subextension of L=K is separable. Equivalently, L=K is separable when every
element of L is separable over K.
This definition makes no sense if L/K is not an algebraic extension since a non-
algebraic extension is not the union of its finite subextensions.
Theorem 1 has a problem in the infinite-degree case: there is no natural trace map.
However, the conditions in Theorems 2 and 3 both make sense for a general L/K. (In the
case of Theorem 2,
we have to drop the specification of K K L as a product of copies of K , and just leave the
statement about the tensor product having no non-zero nilpotent elements.) It is
left to the reader to check for an infinite algebraic extension L/K that the conditions
of Theorems 2 and 3 match Definition.
The conditions in Theorems 2 and 3 both make sense if L/K is not algebraic, so they
could each potentially be used to define separability of a completely arbitrary field
extension. But there is a problem: for transcendental (that is, non-algebraic)
extensions the conditions in Theorems 2 and 3 are no longer equivalent. Indeed,
take L = K(u), with u transcendental over K. Then
K K L = K(u) is a field, so the condition in Theorem 2 is satisfied. However, the zero derivation
on K has more than one extension to K(u): the zero derivation on K(u) and
differentiation with respect to u on K(u).
Definition: A commutative ring with no nonzero nilpotent elements is called reduced.
A domain is reduced, but a more worthwhile example is a product of domains, like
F3 × Q[X], which is not a domain but is reduced.

Definition: An arbitrary field extension L/K is called separable when the ring K K L is reduced.

Using this definition, in characteristic 0 all field extensions are separable. In

characteristic p, any purely transcendental extension is separable. The condition in
Theorem 3, that derivations on the base field admit unique extensions to a larger
field, characterizes not separable field extensions in general, but separable
algebraic field extensions.

A condition equivalent to that in Definition is that F K L is reduced as F runs over

the finite extensions of K.

The condition that K K L is reduced makes sense not just for field extensions L/K, but for any
commutative K-algebra. Define an arbitrary commutative K-algebra A to be separable when
the ring K  A is reduced. This condition is equivalent to A  F being reduced for every finite
K 

extension field F/K.

Example: Let A = K[X]/(f(X)) for any non-constant f(X)  K[X]. The polynomial
f(X) need not be irreducible, so A might not be a field. It is a separable K-algebra
precisely when f(X) is a separable polynomial in K[X].
When [A : K] is finite, an analogue of Theorem 1 can be proved: A is a separable K-
algebra if and only if the trace pairing hx; yi = Tr (xy) from A × A to K is non-
degenerate.
A/K

LOVELY PROFESSIONAL UNIVERSITY

68
Notes Traces

Let A be a finite-dimensional commutative K-algebra (with identity), such as a finite

extension
field of K or the product ring Kn or even a mixture of the two: a product of finite extensions
of K. To any a 2 A we associate the K-linear map m : A  A which is left multiplication by a: 

x  x:
For a; b  A and   K, m = m + m and m = am , so m is a K-linear map.
a+b a b aa a a

Definition: For a finite-dimensional K-algebra A, the trace of a  A is the trace of m .

a

That is, the trace of a is tr(m )  K, usually written as Tr (a), so Tr : A  K.

The trace from A
a A/K A/K
to K is K-linear, hence identically zero or surjective since K is a one-dimensional K-vector space.
Example: Since m is the identity function, Tr (1) = [A : K].
1 A/K

Example: Suppose a  A is nilpotent: ar = 0 for some r  1. Then mr  0, so m is a nil

a a
transformation. Thus its eigenvalues are all 0, so Tr (a) = 0.
A/K

We now consider a finite-dimensional L-algebra A with K a subfield of L such that

[L : K] < . We have finite-dimensional algebras A/L, A/K, and L/K. The next
theorem is called the transitivity of the trace.
Theorem 7: In the above notation, TrA/K = TrL/K TrA/L. In particular, if a  L, then TrA
[A : L]Tr (a).
L/K

Proof: Let (e1; : : : ; em) be an ordered L-basis of A and (f1; : : : ; fn) be an ordered K-basis
of L. Thus as an ordered K-basis of A we can use
(e f ,....., e f ,....., e f ,....., e f ):
11 1n m1 mn

For a  A, let
m n
ae
j

cijei ,cf
ij s

bijrsfr ,

i1 r1

for c  L and b  K. Thus a(ef ) = b e f . So

ij ijrs js i r ijrs ir

[m ] = (c ), [m ] = (b ), [m ] = ([m ] ):
a A/L ij cij L/K ijrs a A/K cij L/K

Thus

Tr (Tr (a)) = Tr ( c )
L/K A/L L/K
 ii

= Tr L /K
(c )
ii

= b
 iirr

i r

= Tr (a).
A/K

Theorem 8: Let A and B be finite-dimensional K-algebras. For (a; b) in the product

ring A × B, Tr (a, b) = Tr (a) + Tr (b).
(A×B)/K A/K B/K

Proof: Let e ,....., e be a K-basis of A and f ,....., f be a K-basis of B. In A × B, e . f = 0.

Therefore,
1 m 1 n i j
the matrix for multiplication by (a, b), with respect to the K-basis { e f }, is a block-diagonal
ij

[ma ] 0
matrix   , whose trace is Tr A/K(a) + TrB/K(b).

 0 [mb ] 
LOVELY PROFESSIONAL UNIVERSITY

69
 Unit 9 : Separable Extensions

Theorem 9: Let A be a finite-dimensional K-algebra, L/K be a field extension, and B = L A be Notes

K

the base extension of A to an L-algebra. For a  A, Tr (1a) = Tr (a).

B/L A/K

Proof: Let e ,....., n

e be a K-basis of A. Write ae = c e , so the matrix for m in this basis is (c ).


1 n j i1 ij i a ij
The tensors 1  e ,....., 1  en are an L-basis of B, and we have
1

n
(1  a) (1  ej) = 1  aej = cij(1  ei ),
i1

so the matrix for mon B is the same as the matrix for m on A. Thus Tr (a) = Tr (1  a).
1a a A/K B/L

Remark: Because m and m have the same matrix representation, not only are their traces the
1a a
same but their characteristic polynomials are the same.
Theorem 10: Let A be a finite-dimensional K-algebra. For any field extension L/K,
the base extension by K of the trace map A  K is the trace map L  A  L. That is,
the function
K

idTr (LKA)/L
: LK A  L which sends an elementary tensor x  a to xTr (a) is the trace map
A/K K

A/K A/K

Tr .
(LKA)/L

Proof: We want to show Tr (t) = (idTr )(t) for all t  L A. The elementary tensors
additively span L A so it succes to check equality when t = x  a for x  K and a  A. This means
K

we need to check Tr (x  a) = xTr (a).

(LKA)/K A/K

Pick a K-basis e1,..., en for A and write aej = i1 cijei with cij  K. The elementary tensors 1  e ,..., 1  e are
an L-basis of LK A and
1 n

n n
(x  a)(1  ej )  x  aej  cij(x  ei )  cijx(1  ei )
i1 i1

by the definition of the L-vector space structure on L A. So the matrix for multiplication by
K
x  a in the basis {1  e } is (c x), which implies
i ij

n n
Tr(LKA)/L (x  a)  ciix  xcii  xTrA /K (a).
i1 i1

Derivations
A derivation is an abstraction of differentiation on polynomials. We want to work
with derivations on fields, but polynomial rings will intervene, so we need to
understand derivations on rings before we focus on fields.
Let R be a commutative ring and M be an R-module (e.g., M = R). A derivation on R with
values in M is a map D : R  M such that D(a + b) = D(a) + D(b) and D(ab) = aD(b) +
bD(a). Easily, by induction D(an) = nan-1D(a) for any n  1. When M = R, we will speak of
a derivation on R.
Example: For any commutative ring A, differentiation with respect to X on A[X] is a
derivation on A[X] (R = M = A[X]).
Example: Let R = A[X] and M = A as an R-module by f(X)a := f(0)a. Then D: R  M
by D(f) = f’(0) is a derivation.

Example: Let D : R  R be a derivation. For f(X)   i aiXi in R[X], set fD(X) =  i Dai X i .
This is the application of D coefficentwise to f(X). The operation f  fD is a
derivation on R[X] (to check the product rule, it suffices to look at monomials).

LOVELY PROFESSIONAL UNIVERSITY

70
 If R = F [u] and D is the usual u-derivative on F [u], then the polynomial f(X) = (u 3 + u)X4 +
Notes uX3
2 2

+ u2X + 1 in R[X] has fD(X) = (u2 + 1)X4 +

X3.
Any element of R satisfying D(a) = 0 is called a D-constant, or just a constant if the
derivation is understood. The constants for a derivation form a subring. For
instance, from the product rule, taking a = b = 1, we obtain D(1) = 0.
Example: The set of all constants for X-differentiation on K[X] is K when K has
characteristic 0 and K[Xp] when K has characteristic p.
Example: If D: R  R is a derivation and f  fD is the corresponding derivation on
R[X], its ring of constants is C[X], where C is the constants for D.
We will generally focus on derivations from R to R, although it will be convenient to
allow R-modules as the target space for derivations in Corollary, which is used in
the main text in the proofs of Theorem 3 and Lemma.
Example: Let’s check that any derivation on K[X] which has the elements of K
among its constants has the form D(f) = hf’ for some h  K[X]. (When h = 1, this is
the usual X-derivative.)
When K is among the constants of D, D is K-linear: D(cf) = cD(f) + fD(c) = cD(f).
Therefore, D is determined by what it does to a K-basis of K[X], such as the power
functions Xn. By induction, D(Xn) = nXn-1D(X) for all n  1. Therefore, by linearity,
D(f) = f’(X)D(X) for every f  K[X]. Set h = D(X).
Theorem 11: Let R be a domain with fraction field K. Any derivation D: R  K uniquely extends
to D : K  K, given by the quotient rule: D (a/b) = (bD(a) – aD(b))/b2.
Proof: Suppose there is an extension of D to a derivation on K. Then, if x = a/b is in K
(with a, b
 A), a = bx, so

D(a) = bD(x) + xD(b):

Therefore in K,

D(a) – xD(b) b
D(x) = 
b

To see, conversely, that this formula does give a derivation D on K, first we check it
is well-defined: if a/b = c/d (with b and d nonzero), then ad = bc, so

aD(d) + dD(a) = bD(c) + cD(b).

Therefore,

bD(a) – aD(b) dD(a) – cD(d) d2 (bD(a) – aD(b)) – b2 (dD(c) –

 =

b2 d2 b2 d2

bd(dD(a) – bD(c)) – d2aD(b) +

= b2d2

bd(cD(b) – aD(d)) – d2aD(b) +

= b2d2

(bc – ad)dD(b) – (ad – bc)bD(d)

=

b2d2
= 0 since ad = bc.

That D satisfies the sum and product rules is left to the reader to check.

LOVELY PROFESSIONAL UNIVERSITY

71
 Unit 9 : Separable Extensions

Theorem 12: Let L/K be a finite extension of fields, and D: K  K be a derivation. Suppose Notes
a  L is separable over K, with minimal polynomial (X)  K[X]. That is, (X) is irreducible in
K[X], () = 0, and ’()  0. Then D has a unique extension from K to a derivation on the field
K(), and it is given by the rule

D(a)
D(f()) = f D
() – f'()
'(a)

for any f(X)  K[X].

Proof: The rule (B.1) looks bizarre at first. To make it less so, we start by assuming
D has an extension to K(), and prove by a direct computation that it must be given
by the indicated
r

formula. For any   K(), write  = f(), where f(X) = c X and c  K. Then

i=0
i i i
r
i i-1 D
D() = D(f()) = (D(ci ) + ci (i D())) = f () + f'()D().
i=0

Taking f(X) = (X) to be the minimal polynomial of  over K, f() = 0, so if D has an

extension to K() then (B.2) becomes
0 = D() + ’()D(),
which proves (since ’()  0) that D() must be given by the formula –D()/’().
Plugging this formula for D(), shows D() must be given by the formula. Since 
was a general element of K(), this proves D has at most one extension to a
derivation on K().
Now, to show the formula works, we start over and define
D
D  ()

D(f()) : = f (a) – f'() .

'()

We need to show this formula is well-defined.

Suppose f () = f () for f ; f  K[X]. Then f (X)  f (X) mod (X), say
1 2 1 2 1 2

f (X) = f (X) + (X)k(X)

1 2

for some k(X)  K[X]. Differentiating both sides with respect to X in the usual way,
f’ (X) = f’ (X) + (X)k’(X) + ’(X)k(X).
1 2

Evaluating at X = ,
f’ () = f’ () + ’()k().
1 2

Since ’()  0, we divide by ’() and multiply through by –D() to get

D (a) D() D

–f' ()  f' (a)   ()k

1 2

'a '()

We want to add f1D() to both sides. First, apply D to the coefficients in (B.3), which
is a derivation on K[X], to get

f1D (X)  f2D (X)  (X)Dk (X)  D (X)k(X).

Therefore,

f1D (a)  f2D (a)  D ()k(a).

LOVELY PROFESSIONAL UNIVERSITY

72
 Notes Add this to both sides to get

D D
 (a)  (a)
f1D ()  f2' ()  f2D ()  D ()k()  f2' ()  D ()k()
'(a) '(a)
D ()
 f2D ()  f2' (a) .
 '()

This proves the formula for a derivation on K() is well-defined. It is left to the
reader to check this really is a derivation.

Example: In contrast with Theorem 12, consider K = F (u) and L = K() where  is a root
p

of Xp – u  K[X]. This is an inseparable irreducible polynomial over K. The u-

derivative on K does not have any extension to a derivation on L. Indeed, suppose
the u-derivative on K has an extension to L, and call it D. Applying D to the equation
ap = u gives
pp-1D() = D(u).
The left side is 0 since we’re in characteristic p. The right side is 1 since D is the u-
derivative on F (u). This is a contradiction, so D does not exist.
p

Corollary: Let L/K be a finite extension of fields. For any derivation D: K  L and 
 L which is separable over K, D has a unique extension to a derivation K()  L. If
D(K)  K then D(K())  K().
Proof: Follow the argument in the proof of Theorem 12, allowing derivations to
have values in L rather than in K(). The formula for D(f()) still turns out to be the
same as in (B.1). In particular, if D(K)  K then the extension of D to a derivation on
K() actually takes values in K().

Self Assessment

1. A ................... ring with no non-zero nilpotent element is called reduced.

(a) associative ring (b) commutative ring
(c) multiplicative ring (d) addition ring

2. An arbitrary field extension ................... is called separable when the ring K u L is

reduced.

(a) L-1k (b)

-1
(c) K/L (d)
3. If L/K be a ................... extensions. Then L is separable over K. If and only if any
derivation of K has a unique extension to a derivative of L.
(a) finite (b) infinite
(c) domain (d) split

4. The extension F2 ( u)/F2 /4). Since F2 ( u)  F2 x/x2  u, which then the non-zero
nilpotent element ...................
(a) X  u (b) u

(c) X1  u (d) u 

LOVELY PROFESSIONAL UNIVERSITY

73
 Unit 9 : Separable Extensions

5. If D : R  R is a derivatives and f  F0 is the corresponding derivation on R[x] from its ring Notes
of constants in C[x], where ................... is the constant for ................... , f(x) =
 a1x1 in R[x]and fd(x)  
u
D(ai )Xi

(a) D, C (b) C, D
(c) X, C (d) C, D

9.2 Summary

 Let L/K be a finite extension. Then L is separable over K if and only if any derivation of K has a unique extension
to a derivation of L.
 If L = K(a ,....., a ) and each a is separable over K then every element of L is separable over
1 r i
K (so L/K is separable).
 Let L/K be a finite extension and F be an intermediate field. If L/F and F/K are separable then L/K is separable.
 The proof of Theorem 2 implies Theorem 4.
 Let L/K be an extension of fields, and   L be algebraic over K. Then is separable over K if and only if any
derivation on K has a unique extension to a derivation on K().
 A commutative ring with no nonzero nilpotent elements is called reduced.
 A domain is reduced, but a more worthwhile example is a product of domains, like F3 × Q[X], which is not a
domain but is reduced.

 An arbitrary field extension L/K is called separable when the ring K K L is reduced.

 A derivation is an abstraction of differentiation on polynomials. We want to work with derivations on fields, but
polynomial rings will intervene, so we need to understand derivations on rings before we focus on fields.

9.3 Keywords

Separability: Separability of a finite field extension L/K can be described in several

different ways.
Commutative Ring: A commutative ring with no nonzero nilpotent elements is called reduced.
Domain: A domain is reduced, but a more worthwhile example is a product of
domains, like F3 × Q[X], which is not a domain but is reduced.
Derivation: A derivation is an abstraction of differentiation on polynomials.

9.4 Review Questions

1. Let R be a domain with fraction field K. Any derivation D: R  K uniquely extends to
D : K  K, given by the quotient rule: D (a/b) = (bD(a) – aD(b))/b2. Prove it.

2. Let L/K be a finite extension of fields, and D: K  K be a derivation. Suppose a  L is separable over K, with
minimal polynomial (X)  K[X]. That is, (X) is irreducible in K[X], () = 0, and ’()  0. Then D has a unique
extension from K to a derivation on the field K(), and it is given by the rule.

LOVELY PROFESSIONAL UNIVERSITY

74
Notes
Answers: Self Assessment

1. (b) 2. (b) 3. (a) 4. (a) 5. (b)

9.5 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

75
Sachin Kaushal, Lovely Professional University Unit 10: Galois
Theory

Unit 10: Galois Theory Notes

CONTENTS
Objectives
Introduction
10 .1 Galois Theory
10.2 Repeated Roots
10.3 The Fundamental Theorem
10.4 Summary
10.5 Keywords
10.6 Review Questions
10.7 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss Galois theory
 Describe repeated roots

Introduction

In the last unit, you have studied about extension field. This unit will provide
information related to Galois theory.

10.1 Galois Theory

This gives the definition of the Galois group and some results that follow
immediately from the definition. We can give the full story for Galois groups of finite
fields.
We use the notation Aut(F) for the group of all automorphisms of F, that is, all one-to-one
functions from F onto F that preserve addition and multiplication. The smallest subfield
containing the identity element 1 is called the prime subfield of F. If F has characteristic
zero, then its prime subfield is isomorphic to Q, and if F has characteristic p, for some
prime number p, then its prime subfield is isomorphic to Z . In either case, for any
automorphisms  of F we must have
p
(x) = x for all elements in the prime subfield of F.
To study solvability by radicals of a polynomial equation f(x) = 0, we let K be the
field generated by the coefficients of f(x), and let F be a splitting field for f(x) over
K. Galois considered permutations of the roots that leave the coefficient field fixed.
The modern approach is to consider the automorphism determined by these
permutations. The first result is that if F is an extension field of K, then the set of all
automorphism  : F  F such that (a) = a for all a  K is a group under
composition of functions. This justifies the following definitions.

LOVELY PROFESSIONAL UNIVERSITY

76
 Notes Definition: Let F be an extension field of K. The set
{   Aut(F) | (a) = a for all a  K }
is called the Galois group of F over K, denoted by Gal(F/K).
Definition: Let K be a field, let f(x)  K[x], and let F be a splitting field for f(x) over
K. Then Gal(F/K) is called the Galois group of f(x) over K, or the Galois group of the
equation f(x) = 0 over K.
Proposition states that if F is an extension field of K, and f(x)  K[x], then any
element of Gal(F/K) defines a permutation of the roots of f(x) that lie in F. The next
theorem is extremely important.
Theorem 1: Let K be a field, let f(x)  K[x] have positive degree, and let F be a splitting
field for
f(x) over K. If no irreducible factor of f(x) has repeated roots, then j Gal(F=K)j = [F : K].
This result can be used to compute the Galois group of any finite extension of any finite
field, but first we need to review the structure of finite fields. If F is a finite field of
characteristic p, then it
is a vector space over its prime subfield Z , and so it has p n elements, where [F : Z ] = n.
The
p p
structure of F is determined by the following theorem.
Theorem 2: If F is a finite field with pn elements, then F is the splitting field of the
polynomial
n

xp – x over the prime subfield of F.

n
The description of the splitting field of xp – x over Z shows that for each prime p and each
p

positive integer n, there exists a field with p n elements. The uniqueness of splitting
fields shows that two finite fields are isomorphic iff they have the same number of
elements. The field with pn elements is called the Galois field of order pn, denoted
by GF(pn). Every finite field is a simple extension of its prime subfield, since the
multiplicative group of nonzero elements is cyclic, and this implies that for each
positive integer n there exists an irreducible polynomial of degree n in Z [x].
p

If F is a field of characteristic p, and n  Z+, then {a  F | apn = a} is a subfield of F, and

this observation actually produces all subfields. In fact, Proposition 6.5.5 has the
following statement: Let F be a field with p n elements. Each subfield of F has p m
elements for some divisor m of n.
Conversely, for each positive divisor m of n there exists a unique subfield of F with pm
elements. If F is a field of characteristic p, consider the function  : F  F defined by (x)
= xp. Since F has characteristic p, we have (a + b) = (a + b)p = ap + bp = (a) + (b),
because in the binomial expansion of (a + b) p each coefficient except those of ap and bp
is zero. (The coefficient (p!)/(k!(p – k)!) contains p in the numerator but not the
denominator since p is prime, and so it must be equal to zero in a field of characteristic
p.) It is clear that  preserves products, and so  is a ring homomorphism. Furthermore,
since it is not the zero mapping, it must be one-to-one. If F is finite, then  must also be
onto, and so in this case  is called the Frobenius automorphism of F.

Note that n(x) = xpn (Inductively, n(x) = (n-1(x))p = (xpn-1 ) p = xpn .) Using an
appropriate power of the Frobenius automorphism, we can prove that the Galois
group of any finite field must be cyclic.
Theorem 3: Let K be a finite field and let F be an extension of K with [F : K] = m.
Then Gal(F/K) is a cyclic group of order m.
Outline of the proof: We start with the observation that F has pn elements, for some
positive

integer n. Then K has pr elements, for r = n/m, and F is the splitting field of x pn x
over its prime subfield, and hence over K. Since f(x) has no repeated roots, to
conclude that |Gal(F/K)|
= m. Now define  : F  F to be the rth power of the Frobenius automorphism. That is,
define

LOVELY PROFESSIONAL UNIVERSITY

77
 Unit 10: Galois Theory

Notes
(x) = xpr . To compute the order of in Gal(F/K), first note that m is the identity since m(x) =

xprm = xpn  x for all x  F. But  cannot have lower degree, since this would give
a polynomial with too many roots. It follows that  is a generator for Gal(F/K).

10.2 Repeated Roots

In computing the Galois group of a polynomial, it is important to know whether or
not it has repeated roots. A field F is called perfect if no irreducible polynomial over
F has repeated roots. This section includes the results that any field of
characteristic zero is perfect, and that any finite field is perfect.
In the previous section, we showed that the order of the Galois group of a polynomial
with no repeated roots is equal to the degree of its splitting field over the base field. The
first thing in this section is to develop methods to determine whether or not a polynomial
has repeated roots.
Let f(x) be a polynomial in K[x], and let F be a splitting field for f(x) over K. If f(x) has the
factorization f(x) = (x – r )m ... (x – r )m
1 over F, then we say that the root r has multiplicity m .
t
1 t i i
If m = 1, then r is called a simple root.
i i

Let f(x)  K[x], with f(x) = t a xk . The formal derivative f’(x) of f(x) is defined by the formula
k=0 k

f'(x) = t ka xk-1 , where ka denotes the sum of a added to itself k times. It is not difficult to
k
k=0 k k

show from this definition that the standard differentiation formulas hold. Proposition
shows that the polynomial f(x)  K[x] has no multiple roots iff it is relatively prime
to its formal derivative f’(x). Proposition shows that f(x) has no multiple roots unless
char(K) = p  0 and f(x)
has the form f(x) = a + a xp + a x2p + ... + a xnp.
0 1 2 n

A polynomial f(x) over the field K is called separable if its irreducible factors have
only simple roots. An algebraic extension field F of K is called separable over K if
the minimal polynomial of each element of F is separable. The field F is called
perfect if every polynomial over F is separable.
Theorem states that any field of characteristic zero is perfect, and a field of
characteristic p > 0 is perfect if and only if each of its elements has a pth root in
the field. It follows immediately from the theorem that any finite field is perfect.
To give an example of a field that is not perfect, let p be a prime number, and let K = Z . Then in
p

the field K(x) of rational functions over K, the element x has no pth root. Therefore,
this rational function field is not perfect.
The extension field F of K is called a simple extension if there exists an element u 
F such that F = K(u). In this case, u is called a primitive element. Note that if F is a
finite field, then Theorem shows that the multiplicative group F x is cyclic. If the
generator of this group is a, then it is easy to see that F = K(a) for any subfield K.
Theorem shows that any finite separable extension is a simple extension.

10.3 The Fundamental Theorem

Here we study the connection between subgroups of Gal(F/K) and fields between K
and F. This is a critical step in proving that a polynomial is solvable by radicals if
and only if its Galois group is solvable.

LOVELY PROFESSIONAL UNIVERSITY

78
 Notes Let F be a field, and let G be a subgroup of Aut(F). Then
{a  F | (a) = a for all   G}
is called the G-fixed subfield of F, or the G-invariant subfield of F, and is denoted by
FG. (Proposition shows that F G is actually a subfield of F.) If F is the splitting field
over K of a separable polynomial and G = Gal(F/K), then Proposition shows that F G
= K. Artin’s lemma provides the first really significant result of the section. It states
that if G is a finite group of automorphism of the field F, and K = F G, then [F : K]  |
G|.
Let F be an algebraic extension of the field K. Then F is said to be a normal extension of
K if every irreducible polynomial in K[x] that contains a root in F is a product of linear
factors in F[x]. With this definition, the following theorem and its corollary can be proved
from previous results.
Theorem 4: The following are equivalent for an extension field F of K:
(1) F is the splitting field over K of a separable polynomial;
(2) K = FG for some finite group G of automorphism of F;
(3) F is a finite, normal, separable extension of K.
As a corollary, we obtain the fact that if F is an extension field of K such that K = F G
for some finite group G of automorphisms of F, then G = Gal(F/K).
The next theorem is the centerpiece of Galois theory. In the context of the fundamental
theorem,
we say that two intermediate subfields E and E are conjugate if there exists   Gal(F/K)
such
1 2
that (E ) = E . Proposition states that if F is the splitting field of a separable polynomial over K,
1 2
and K  E  F, with H = Gal(F/E), then Gal(F/(E)) = H-1, for any   Gal(F/K).
Theorem 5 (The Fundamental Theorem of Galois Theory): Let F be the
splitting field of a separable polynomial over the field K, and let G = Gal(F/K).
(a) There is a one-to-one order-reversing correspondence between subgroups of G and
subfields of F that contain K:
(i) The subfield FH corresponds to the subgroup H, and H = Gal(F/F H).
(ii) If K  E  F, then the corresponding subgroup is Gal(F/E), and E = F Gal(F/E).
(b) [F : FH] = |H| and [FH : K] = [G : H], for any subgroup H of G.
(c) Under the above correspondence, the subgroup H is normal iff F H is a normal
extension of K. In this case, Gal(E/K) Gal(F/K) / Gal(F/E).
For example, suppose that F is a finite field of characteristic p, and has p m
elements. Then [F : GF(p)] = m, and so G = Gal(F= GF(p)) is a cyclic group of
degree m by Corollary. Since G is cyclic, the subgroups of G are in one-to-one
correspondence with the positive divisors of m. Proposition shows that the subfields
of F are also in one-to-one correspondence with the positive divisors of m.
Remember that the smaller the subfield, the more automorphisms will leave it
invariant. By the Fundamental Theorem of Galois Theory, a subfield E with [E :
GF(p)] = k corresponds to the cyclic subgroup with index k, not to the cyclic
subgroup of order k.

Self Assessment

1. If F has characteristics zero, then its prime subfield is isomorphic to Q and if F has
characteristics P, for some prime number P, then its prime subfield is ................ to Zp.
(a) homomorphic (b) isomorphic
(c) automorphism (d) polynomial

LOVELY PROFESSIONAL UNIVERSITY

79
 Unit 10: Galois
Theory

2. Let F be extension field of K. The set { Q  Aut(F) | Q(a) = a for all a  K } is

Galois group is denoted by ................ Notes

(a) Gal(F/K) (b) Gal(u/F)

(c) Gal-1(K/F) (d) Gal(k × F)
3. Let K be a finite field and let F be an extension of K with [F : k] = m. Then
Gal(F/k) is a
................ group of order m.
(a) cyclic (b) polynomial
(c) permutation (d) finite
4. A polynomial f(x) over the field k is called ................ if its irreducible factors
have only simple roots.
(a) spittery field (b) extension field
(c) separable (d) finite field
5. The ................ F of K is called simple extensions. If then exist an element u 
F. Such that F = K(u).
(a) finite field (b) extension field
(c) separable field (d) spliting field

10.4 Summary
 Let F be an extension field of K. The set
{   Aut(F) | (a) = a for all a  K }

is called the Galois group of F over K, denoted by Gal(F/K).

 Let K be a field, let f(x)  K[x], and let F be a splitting field for f(x) over K.
Then Gal(F/K) is called the Galois group of f(x) over K, or the Galois group of the
equation f(x) = 0 over K.
 It states that if F is an extension field of K, and f(x)  K[x], then any element
of Gal(F/K) defines a permutation of the roots of f(x) that lie in F. The next theorem
is extremely important.
 Let K be a field, let f(x)  K[x] have positive degree, and let F be a splitting
field for f(x) over K. If no irreducible factor of f(x) has repeated roots, then j
Gal(F=K)j = [F : K].
This result can be used to compute the Galois group of any finite extension of any
finite field, but first we need to review the structure of finite fields. If F is a finite
field of characteristic p, then it is a vector space over its prime subfield Z , and so it
has pn elements,
p
where [F : Z ] = n. The structure of F is determined by the following theorem.
p

 If F is a finite field with pn elements, then F is the splitting field of the

polynomial xpn – x
over the prime subfield of F.
 Let K be a finite field and let F be an extension of K with [F : K] = m. Then
Gal(F/K) is a cyclic group of order m.
 (The fundamental theorem of Galois theory) Let F be the splitting field of a
separable polynomial over the field K, and let G = Gal(F/K).

LOVELY PROFESSIONAL UNIVERSITY

80
 Notes 10
.5 Keyword
Prime Subfield: If F is a finite field with pn elements, then F is the splitting field of the
polynomial
n

xp – x over the prime subfield of F.

The Fundamental Theorem of Galois Theory: Let F be the splitting field of a

separable polynomial over the field K, and let G = Gal(F/K).

10.6 Review Questions

1. Determine the group of all automorphisms of a field with 4 elements.

2. Let F be the splitting field in C of x4 + 1.

(a) Show that [F : Q] = 4.

(b) Find automorphisms of F that have fixed fields Q( 2 ), Q(i), and Q( 2 i),
respectively.
3. Find the Galois group over Q of the polynomial x4 + 4.
4. Find the Galois groups of x3 – 2 over the fields Z and Z .
5 11

5. Find the Galois group of x4 – 1 over the field Z .

7

6. Find the Galois group of x3 – 2 over the field Z .

7

7. Let f(x) 2 Q[x] be irreducible over Q, and let F be the splitting field for f(x) over Q. If
[F : Q] is odd, prove that all of the roots of f(x) are real.

8. Find an element  with Q( 2 , i) = Q().

9. Find the Galois group of x6 – 1 over Z .
7

G
10. Prove that if F is a field and K = F for a finite group G of automorphisms of F, then
there are only finitely many subfields between F and K.
11. Let F be the splitting field over K of a separable polynomial. Prove that if Gal(F/K) is
cyclic, then for each divisor d of [F : K] there is exactly one field E with K  E  F and [E : K] = d.
12. Let F be a finite, normal extension of Q for which |Gal(F=Q)| = 8 and each element of
Gal(F/Q) has order 2. Find the number of subfields of F that have degree 4 over Q.
13. Let F be a finite, normal, separable extension of the field K. Suppose that the Galois
group
Gal(F/K) is isomorphic to D . Find the number of distinct subfields between F and K. How
7
many of these are normal extensions of K?

14. Show that F = Q(i, 2 ) is normal over Q; find its Galois group over Q, and find all
intermediate fields between Q and F.

3
15. Let F = Q( 2, 2 ). Find [F : Q] and prove that F is not normal over Q.

16. Find the order of the Galois group of x5 – 2 over Q.

Answers: Self Assessment

1. (b) 2. (a) 3. (a) 4. (c) 5. (b)

LOVELY PROFESSIONAL UNIVERSITY

81
 Unit 10: Galois Theory

10.7 Further Readings Notes

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant
Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

82
 Richa Nandra, Lovely Professional University

Notes Unit 11: Computing Galois Groups

CONTENTS
Objectives

Introduction

11.1 Transitively Group

11.2 Separable Polynomial

11.3 Summary

11.4 Keywords

11.5 Review Questions

11.6 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss transitively and transitive group
 Explain computing Galois group

Introduction

In the last unit, you have studied about Galois theory. In this unit, you will get
information related to computing the Galois groups.

11.1 Transitively Group

Definition: Let G be a group acting on a set S. We say that G acts transitively on

S if for each pair of elements x,y in S there exist an element g in G such that y =
gx.
If G is a subgroup of the symmetric group S , then G is called a transitive group if it
acts
n
transitively on the set { 1, 2, ... , n }.

11.2 Separable Polynomial

Proposition: Let f(x) be a separable polynomial over the field K, with roots r , ... , r in its
1 n

splitting field F. Then f(x) is irreducible over K if and only if Gal(F/K) acts transitively
on the roots of f(x).
Lemma: Let p be a prime number, and let G be a transitive subgroup of S . Then any
nontrivial
p
normal subgroup of G is also transitive.
Lemma: Let p be a prime number, and let G be a solvable, transitive subgroup of S . Then
G
p
contains a cycle of length p.
Proposition: Let p be a prime number, and let G be a solvable, transitive subgroup of S . Then
G
p
is a subgroup of the normalizer in S of a cyclic subgroup of order p.
p

LOVELY PROFESSIONAL UNIVERSITY

83
 Unit 11: Computing Galois
Groups

Let f(x) be a polynomial of degree n over the field K, and assume that f(x) has roots r , r , ... , r in Notes
1 2 n

its splitting field F. The element  of F defined by

 = (r - r ) 2,
i j

where the product is taken over all i, j with 1  i < j  n, is called the discriminant of f(x).
It can be shown that the discriminant of any polynomial f(x) can be expressed as a
polynomial in the coefficients of f(x), with integer coefficients. This requires use of
elementary symmetric functions, and lies beyond the scope of what we have
chosen to cover in the book.
We have the following properties of the discriminant:
(i)   0 if and only if f(x) has distinct roots;
(ii)  belongs to K;
(iii) If   0, then a permutation in S is even if and only if it leaves unchanged the sign of
n

 1  i < j  n(ri - rj )


Proposition: Let f(x) be a separable polynomial over the field K, with discriminant ,
and let F be its splitting field over K. Then every permutation in Gal(F/K) is even if
and only if is the square of some element in K.
We now restrict our attention to polynomials with rational coefficients. The next
lemma shows that in computing Galois groups it is enough to consider polynomials
with integer coefficients. Then a powerful technique is to reduce the integer
coefficients modulo a prime and consider the Galois group of the reduced equation
over the field GF(p).
Lemma: Let f(x) = xn + a xn-1 + · · · + a x + a be a polynomial in Q[x], and assume that
n-1 1 0
a = b / d for d, b , b , ... , b in Z.
i i 0 1 n-1

Then dn f(x/d) is monic with integer coefficients, and has the same splitting field over Q as f(x).
If p is a prime number, we have the natural mapping : Z[x] > Z [x] which reduces each coefficient
p
modulo p. We will use the notation p(f(x)) = f (x).
p

Theorem [Dedekind]: Let f(x) be a monic polynomial of degree n, with integer coefficients and
Galois group G over Q, and let p be a prime such that f (x) has distinct roots. If f (x) factors in Z [x]
p p p
as a product of irreducible factors of degrees n , n , ... , n , then G contains a permutation with the
1 2 k
cycle decomposition
(1,2, ... ,n ) (n +1, n +2, ... , n +n ) · · · (n-n +1, ... ,n),
1 1 1 1 2 k

relative to a suitable ordering of the roots.

Self Assessment

1. IF G is a .................. of symmetric group sn then G is called transitive group. If it acts transitively on Set {1, 2, 3,
n}
(a) sub group (b) cyclic group
(c) permutation group (d) finite group
2. Let P be a prime number and let G be a transitive subgroup of Sp. Then any ..................
normal subgroup of G is also transitive.
(a) trivial (b) non-trivial
(c) finite (d) infinite

LOVELY PROFESSIONAL UNIVERSITY

84
 Notes 3. Let P be a prime number and G be a solvable, transitive subgroup of S . Then G is a
p

subgroup of the normalizer in S of a cyclic subgroup of order ..................

p

(a) P (b) G
(c) S (d) S
p

If f(x) be a polynomial of degree n over the field k and assume that f(x) has roots r ,
4. r ,...r
1 2n

in its splitting field F. Then element  of F defined by

(a)  = (r – r )2 (b)  = 2(r – r )2
1 3 1 j

(c)  = (r – r )-2 (d)  = 3(r – r )3

i j 1 j

11.3 Summary
 Let f(x) be a separable polynomial over the field K, with roots r , ... , r in its splitting
field
1n
F. Then f(x) is irreducible over K if and only if Gal(F/K) acts transitively on the roots of
f(x).
 Let p be a prime number, and let G be a transitive subgroup of S . Then any non-
trivial
p
normal subgroup of G is also transitive.
 Let p be a prime number, and let G be a solvable, transitive subgroup of S . Then G
p
contains a cycle of length p.
 Let p be a prime number, and let G be a solvable, transitive subgroup of S . Then G is
a
p
subgroup of the normalizer in S of a cyclic subgroup of order p.
p

 Let f(x) = xn + a xn-1 + · · · + a x + a be a polynomial in Q[x], and assume that

n-1

a = b / d for d, b , b , ... , b
i i 0 1

 Then dn f(x/d) is monic with integer coefficients, and has the same splitting field over
Q as f(x).
 If p is a prime number, we have the natural mapping : Z[x] > Z [x] which reduces
each
p
coefficient modulo p. We will use the notation p(f(x)) = f (x).
p

 Let f(x) be a monic polynomial of degree n, with integer coefficients and Galois group
G
over Q, and let p be a prime such that f (x) has distinct roots. If f (x) factors in Z [x]
as a
p p p
product of irreducible factors of degrees n , n , ... , n , then G contains a permutation with
1 2 k

the cycle decomposition

(1,2, ... ,n1) (n1+1, n1+2, ... , n1 +n2) · · · (n-n

relative to a suitable ordering of the roots.

11.4 Keywords
Transitive Group: If G is a subgroup of the symmetric group S , then G is called a
transitive
n
group if it acts transitively on the set { 1, 2, ... , n }.
Separable Polynomial: Let f(x) be a separable polynomial over the field K, with roots
r , ... , r
1 n

in its splitting field F. Then f(x) is irreducible over K if and only if Gal(F/K) acts
transitively on the roots of f(x).
LOVELY PROFESSIONAL UNIVERSITY

85
 Unit 11: Computing Galois
Groups

Notes
11.5 Review Questions
1. Give the order and describe a generator of the Galois group of GF (729) over GF(9).
2. Determine the Galois group of each of the following polynomials in Q[x]; hence, determine the solvability of
each of the polynomials

(a) x5 – 12x2 + 2 (b) x5 – 4x4 + 2x + 2

(c) x3 –5 (d) x4 – x2 – 6
(e) x5 +1 (f) (x2 – 2) (x2 + 2)

(g) x8 –1 (h) x8 +1
4 2
(i) x – 3x – 10
3. Find a primitive element in the splitting field of each of the following polynomials in Q[x].

(a) x4 –1 (b) x4 – 2x2 – 15

(c) x4 – 8x2 + 15 (d) x3 –2
4. Prove that the Galois group of an irreducible quadratic polynomial is isomorphic to Z .
2

5. Prove that the Galois group of an irreducible cubic polynomial is isomorphic to S or Z .

3 3

Answers: Self Assessment

1. (a) 2. (b) 3. (a) 4. (a)

11.6 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

86
 Richa Nandra, Lovely Professional University

Notes Unit 12: Invariant Subfield

CONTENTS
Objectives
Introduction
12.1 G-invariant Subfield
12.2 Summary
12.3 Keywords
12.4 Review Questions
12.5 Further Readings

Objectives

After studying this unit, you will be able to:

 Define G-invariant subfield
 Discuss examples related to subfield

Introduction

In the last unit, you have studied about computing Galois theory and groups. In this
unit, you will get information related to fundamental theorem.

12.1 G-invariant Subfield

Proposition: Let F be a field, and let G be a subgroup of Aut(F). Then

{ a in F |  (a) = a for all  in G }
is a subfield of F.
Definition: Let F be a field, and let G be a subgroup of Aut (F). Then
{ a in F |  (a) = a for all  in G }
is called the G-fixed subfield of F, or the G-invariant subfield of F, and is denoted by
F G.

Proposition: If F is the splitting field over K of a separable polynomial and G =

Gal(F/K), then FG=K.
Lemma [Artin]: Let G be a finite group of automorphisms of the field F, and let K = F G.
Then
[F:K] |G|.
Let F be an algebraic extension of the field K. Then F is said to be a normal extension of
K if every irreducible polynomial in K[x] that contains a root in F is a product of linear
factors in F[x].

LOVELY PROFESSIONAL UNIVERSITY

87
 Unit 12: Invariant Subfield

The following conditions are equivalent for an extension field F of K: Notes

(1) F is the splitting field over K of a separable polynomial;

(2) K = FG for some finite group G of automorphisms of F;

(3) F is a finite, normal, separable extension of K.

If F is an extension field of K such that K = F G for some finite group G of

automorphisms of F, then G = Gal(F/K).

Example: The Galois group of GF(pn) over GF(p) is cyclic of order n,

generated by the automorphism  defined by  (x) = xp, for all x in GF(p n). This
automorphism is usually known as the Frobenius automorphism of GF(pn).
Let F be the splitting field of a separable polynomial over the field K, and let G = Gal(F/K).
(a) There is a one-to-one order-reversing correspondence between subgroups of G and subfields of F that contain K:
(i) If H is a subgroup of G, then the corresponding subfield is F H, and H = Gal(F/FH).
(ii) If E is a subfield of F that contains K, then the corresponding subgroup of G is H = Gal(F/E), and
E=FH.
(b) For any subgroup H of G, we have

[F : FH] = | H| and [FH : K] = [G : H].

(c) Under the above correspondence, the subgroup H is normal if and only if the subfield E = F H is a normal
extension of K. In this case,
Gal(E/K) Gal(F/K)/Gal(F/E).

In the statement of the fundamental theorem we could have simply said that
normal subgroups correspond to normal extensions. In the proof we noted that if E
is a normal extension of K, then (E)  E for  all in Gal(F/K). In the context of the
fundamental theorem, we say that two intermediate subfields E and E are
conjugate if there exists in Gal(F/K) such that (E ) = E .
1 2 1 2
The next result shows that the subfields conjugate to an intermediate subfield E
correspond to the subgroups conjugate to Gal(F/E). Thus E is a normal extension if
and only if it is conjugate only to itself.
Let F be the splitting field of a separable polynomial over the field K, and let E be a
subfield such that K  E  F, with H = Gal(F/E). If  is in Gal(F/K), then
Gal(F/(E)) = H -1.
[Fundamental Theorem of Algebra]: Any polynomial in C[x] has a root in C.

Example: Prove that if F is a field extension of K and K = F G for a finite

group G of automorphisms of F, then there are only finitely many subfields between
F and K.
Solution: The given condition is equivalent to the condition that F is the splitting
field over K of a separable polynomial. Since we must have G = Gal (F/K), the
fundamental theorem of Galois theory implies that the subfields between F and K
are in one-to-one correspondence with the subgroups of F. Because G is a finite
group, it has only finitely many subgroups.

LOVELY PROFESSIONAL UNIVERSITY

88
 Notes
Example: Let F be the splitting field over K of a separable polynomial. Prove
that if Gal (F/K) is cyclic, then for each divisor d of [F:K] there is exactly one field E
with K E F and [E:K] = d.
Solution: By assumption we are in the situation of the fundamental theorem of
Galois theory, so that there is a one-to-one order-reversing correspondence
between subfields of F that contain K and subgroups of G = Gal (F/K). Because G is
cyclic of order [F:K], there is a one-to-one correspondence between subgroups of G
and divisors of [F:K]. Thus for each divisor d of [F:K] there is a unique subgroup H of
index d. By the fundamental theorem, [F H: K] = [G:H], and so E = F^H is the unique
subfield with [E:K] = d.
Comment: Pay careful attention to the fact that the correspondence between
subfields and subgroups reverses the order

Example: Let F be a finite, normal extension of Q for which | Gal (F/Q) | = 8

and each element of Gal (F/Q) has order 2. Find the number of subfields of F that
have degree 4 over Q.
Solution: Since F has characteristic zero, the extension is automatically separable,
and so the fundamental theorem of Galois theory can be applied. Any subfield E of
F must contain Q, its prime subfield, and then [E:Q] = 4 iff [F:E] = 2, since [F:Q] =
8. Thus the subfields of F that have degree 4 over Q correspond to the subgroups of
Gal (F/Q) that have order 2. Because each nontrivial element has order 2 there are
precisely 7 such subgroups.

Example: Let F be a finite, normal, separable extension of the field K. Suppose that
the
Galois group Gal (F/K) is isomorphic to D . Find the number of distinct subfields between F and
7
K. How many of these are normal extensions of K?
Solution: The fundamental theorem of Galois theory converts this question into the
question of
enumerating the subgroups of D , and determining which are normal. If we use the usual
7
-1
description of D via generators a of order 7 and b of order 2, with ba = a b, then a generates a
7
i
subgroup of order 7, while each element of the form a b generates a subgroup of order 2,
for
0  i < 7. Thus there are 8 proper nontrivial subgroups of D , and the only one that is
normal is
7
< a >, since it has |D | / 2 elements. As you should recall from the description of the conjugacy
7
classes of D conjugating one of the 2-element subgroups by a produces a different subgroup,
7
showing that none of them are normal.

Example: Show that F = Q ( 2,i) is normal over Q; find its Galois group over Q,
and find all intermediate fields between Q and F.
Solution: It is clear that F is the splitting field over Q of the polynomial (x 2 + 1)(x2 –
2), and this polynomial is certainly separable. Thus, F is a normal extension of Q.
It follows that the Galois group is isomorphic to Z × Z . Since the Galois group has 3 proper
2 2
nontrivial subgroups, there will be 3 intermediate subfields E with Q  E  F.
The existence of 3 nontrivial elements begins with the splitting field of x 4+1 over Q.
Comment: Recall that Z is the splitting field of x7 – x = x(x6 – 1).
7

Self Assessment

1. If F is field and G be a subgroup of Aut(F). Then {a in F | Q(a) = a  Q in G} is

called ...............
of F.

LOVELY PROFESSIONAL UNIVERSITY

89
 Unit 12: Invariant Subfield

(a) G-invariant subfield (b) variant subfield Notes

(c) finite field (d) domain subfield
2. G-invariant subfield is denoted by ...............

(a) GF (b) FG

(c) F.G (d) GF-1

3. G be a finite group of automorphisms of the field F and let K = FG then [F : K] ...............|G|.
(a)  (b) 
(c) = (d) 

4. For any subgroup H of G, we have ............... and [F H : K] = [G : H]

(a) [F : FH] = |H| (b) [HF : F]  |H|

(c) [F–1 : H] = |H| (d) [F–1 : H–1]  |H|

5. Let F be an algebraic extensions of the field K. Then F is said to be a ............... of K. If every irreducible polynomial in
K[x] that contains a root in F is a product of linear factors in F[x]
(a) normal extension (b) finite extension
(c) infinite extension (d) subgroup extension

12.2 Summary

 The following conditions are equivalent for an extension field F of K:

(1) F is the splitting field over K of a separable polynomial;

(2) K = FG for some finite group G of automorphisms of F;

(3) F is a finite, normal, separable extension of K.

 If F is an extension field of K such that K = FG for some finite group G of automorphisms of F, then G = Gal(F/K).
 Let F be the splitting field of a separable polynomial over the field K, and let G = Gal(F/K).
(a) There is a one-to-one order-reversing correspondence between subgroups of G and subfields of F that
contain K:
(i) If H is a subgroup of G, then the corresponding subfield is F H, and
H = Gal(F/FH).
(ii) If E is a subfield of F that contains K, then the corresponding subgroup of G is H = Gal(F/E), and
E = FH.
(b) For any subgroup H of G, we have

[F : FH] = | H| and [FH : K] = [G : H].

(c) Under the above correspondence, the subgroup H is normal if and only if the subfield E = F H is a normal
extension of K. In this case,
Gal(E/K) Gal(F/K) / Gal(F/E).

LOVELY PROFESSIONAL UNIVERSITY

90
 Notes  Let F be the splitting field of a separable polynomial over the field K, and let E be a
subfield such that K  E  F, with H = Gal(F/E). If  is in Gal(F/K), then
Gal(F/(E)) = H -1.
 [Fundamental Theorem of Algebra] Any polynomial in C[x] has a root in C.

12.3 Keywords

Normal Extension: Let F be an algebraic extension of the field K. Then F is said to

be a normal extension of K if every irreducible polynomial in K[x] that contains a
root in F is a product of linear factors in F[x].
Frobenius Automorphism: The Galois group of GF(p n) over GF(p) is cyclic of
order n, generated by the automorphism  defined by  (x) = xp, for all x in GF(p n).
This automorphism is usually known as the Frobenius automorphism of GF(p n).

12.4 Review Questions

1. Compute each of the following Galois groups. Which of these field extensions are
normal field extensions? If the extension is not normal, find a normal extension of Q in which the extension field
is contained.

(a) G(Q( 30)/Q) (b) G(Q(4 5)/Q)

3
(c) G(Q( 2, 3, 5)/Q) (d) G(Q( 2, 2,i)

(e) G(Q( 6,i)/Q)

2. Let F K  E be field. If E is a normal extension of F, show that E must also be a

normal extension of K.
3. Let G be the Galois group of a polynomial of degree n. Prove that |G| divides n!.
4. Let F  E. If f(x) is solvable over F, show that f(x) is also solvable over E.

Answers: Self Assessment

1. (a) 2. (b) 3. (b) 4. (a) 5. (a)

12.5 Further Readings

Books
Dan Saracino: Abstract Algebra; A First Course.
Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links
www.jmilne.org/math/CourseNotes/
www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

91
Richa Nandra, Lovely Professional University Unit 13: The Galois Group of a Polynomial

Unit 13: The Galois Group of a Polynomial Notes

CONTENTS
Objectives
Introduction
13.1 Galois Group of Polynomial
13.2 Fundamental Theorem of Symmetric Functions
13.3 Symmetric Polynomial
13.4 Constructible Polygon
13.5 Connection to Pascal’s Triangle
13.6 Summary
13.7 Keywords
13.8 Review Questions
13.9 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss the Galois group of polynomial
 Explain the theorem of Galois theory

Introduction

To study solvability by radicals of a polynomial equation f(x) = 0, we let K be the

field generated by the coefficients of f(x), and let F be a splitting field for f(x) over
K. Galois considered permutations of the roots that leave the coefficient field fixed.
The modern approach is to consider the automorphisms determined by these
permutations. We note that any automorphism of a field F must leave its prime
subfield fixed.

13.1 Galois Group of Polynomial

Proposition: Let F be an extension field of K. The set of all automorphisms  : F >

F such that (a) = a for all a in K is a group under composition of functions.
Definition: Let F be an extension field of K. The set
{ in Aut(F) | (a) = a for all a in K }
is called the Galois group of F over K, denoted by Gal(F/K).
Definition: Let K be a field, let f(x) be a polynomial in K[x], and let F be a splitting
field for f(x) over K. Then Gal(F/K) is called the Galois group of f(x) over K, or the
Galois group of the equation f(x) = 0 over K.

LOVELY PROFESSIONAL UNIVERSITY

92
 Notes Proposition: Let F be an extension field of K, and let f(x) be a polynomial in K[x]. Then any
element of Gal(F/K) defines a permutation of the roots of f(x) that lie in F.
Let f(x) be a polynomial in K[x] with no repeated roots and let F be a splitting field for f(x) ove
K. If  : K > L is a field isomorphism that maps f(x) to g(x) in L[x] and E is a splitting field for g(x
over L, then there exist exactly [F:K] isomorphisms : F -> E such that (a) = (a) for all a in K.
Theorem: Let K be a field, let f(x) be a polynomial in K[x], and let F be a splitting field for f(x)
over K. If f(x) has no repeated roots, then |Gal(F/K)| = [F:K].
Corollary: Let K be a finite field and let F be an extension of K with [F:K] = m. Then Gal
(F/K) is a cyclic group of order m.
If we take K = Zp, where p is a prime number, and F is an extension of degree m, then the
generator of the cyclic group Gal(F/K) is the automorphism  : F -> F defined by (x) = xp, for a
x in F. This automorphism is called the Frobenius automorphism of
F.
A symmetric function on n variables x1,..., xn is a function that is unchanged by any
permutation
of its variables. In most contexts, the term “symmetric function” refers to a polynomial on n
variables with this feature (more properly called a “symmetric polynomial”). Another type of
symmetric functions is symmetric rational functions, which are the rational functions that are
unchanged by permutation of variables.
The symmetric polynomials (respectively, symmetric rational functions) can be expressed as
polynomials (respectively, rational functions) in the elementary symmetric polynomials. This
is called the fundamental theorem of symmetric functions.
A function f(x) is sometimes said to be symmetric about the y-axis if f(–x) = f(x). Examples of
such
functions include |x| (the absolute value) and x2 (the parabola).
13.2 Fundamental Theorem of Symmetric
Functions

Any symmetric polynomial (respectively, symmetric rational function) can be expressed as a

polynomial (respectively, rational function) in the elementary symmetric polynomials on those
variables.
There is a generalization of this theorem to polynomial invariants of permutation groups G,
which states that any polynomial invariant f  R [X1,... Xn] can be represented as a finite linea
combination of special G-invariant orbit polynomials with symmetric functions as coefficients,
i.e.,

f   p1 (1 ,...,n ) orbit G(t),

r special

where p R[X,...,X],
1 1 n

and  , ...,  are elementary symmetric functions, and t = Xe , ..., Xe 1 n are special terms
1 n 1 n

Furthermore, any special term t has a total degree  n(n – 1)/2, and a maximal
variable degree  n – 1.

13.3 Symmetric Polynomial

A symmetric polynomial on n variables x ,..., x (also called a totally symmetric polynomial)
is
1 n

a function that is unchanged by any permutation of its variables. In other words,

the symmetric polynomials satisfy
f(y , y , ..., y ) = f(x , x ,..., x ),
1 2 n 1 2 n
where y = x and  being an arbitrary permutation of the indices 1, 2, ..., n.
i (i)

LOVELY PROFESSIONAL UNIVERSITY

93
 Unit 13 : The Galois Group of a Polynomial

For fixed n, the set of all symmetric polynomials in n variables forms an algebra of Notes
dimension n. The coefficients of a univariate polynomial f(x) of degree n are algebraically
independent symmetric polynomials in the roots of f, and thus form a basis for the set of all such
symmetric polynomials.
There are four common homogeneous bases for the symmetric polynomials, each of which is
indexed by a partition  (Dumitriu et al., 2004). Letting l be the length of , the elementary
functions e , complete homogeneous functions h , and power-sum functions p are defined for
  
l = 1 by

e =  ...(2)
1

j1  j2 ... j 1

...(3)

h 1
= 
...(4)
m1 ...mn l1 j1

p = x

 .
1

j1
...(5)

and for l > 1 by

l
s = s
   i
...(6)
i1

where s is one of e, h or p. In addition, the monomial functions m are

defined as

x 1 2

m =   (1) s
x
(2)
S

where S is the set of permutations giving distinct terms in the sum and  is considered to be


infinite.
As several different abbreviations and conventions are in common use, care must
be taken when determining which symmetric polynomial is in use.
The elementary symmetric polynomials  (x , ..., x ) (sometimes or e ) on n variables
denoted 


k 1 n k

{x , ..., x } are defined by ...(7)

1 n

( , ..., x
 1x xn)
1
=  i ...(8)
1in

( , ..., xx ...(9)
 2x xn)
1
=  i j

1i jn

( , ..., ...(10)
x xx
 3x xn)
1
=  i j k
1i jk
n

( , ...,
 4x xn)
1
=  xixjxk xl ...(12)
1i jk ln

( , ..., x
 5x xn)
1
=  i

1in
The kth elementary symmetric polynomial is implemented in Mathematica as Symmetric
Polynomial [k, {x , ..., x }]. Symmetric Reduction [f, {x , ..., x }] gives a pair of polynomials
1 n 1 n
{p, q} in x , ..., x where is the symmetric part and is the remainder.
1 n
LOVELY PROFESSIONAL UNIVERSITY

94
 Notes Alternatively,  (x ,..., x ) can be defined as the coefficient of xn-j in the generating functio
j 1 n

 (x  xi ). ...(13
1in

For example, on four variables x , ..., x , the elementary symmetric polynomials are
1 4

 (x , x , x , x ) = x + x + x+ x
1 1 2 3 4 1 2 3 4

 (x , x , x , x ) = x x +x x+ x x+x x +
2 1 2 3 4 1 2 1 3 1 4 2 3

 (x , x , x , x ) = x x x + xx x +x x x +
3 1 2 3 4 1 2 3 1 2 4 1 3 4

 (x , x , x , x ) = x x x x
4 1 2 3 4 1 2 3 4
The power sum S (x ,..., x ) is defined by
p 1 n

Sp(x1, ..., xn) = xkp .

k 1

The relationship between * and  ,..., is given by the so-called Newton-Girard formulas.
The
1 p
related function s ( , ...,  ) with arguments given by the elementary symmetric polynomials
p 1 n
(not x ) is defined by
n

s ( ,..., ) = (–1)p–1 S (x ,...,x ) ...(19)

p 1 n p 1 n

n
...(20)
= (1)p1 xkp .

k 1

It turns out that s ( , ..., ) is given by the coefficients of the generating function
p1 n
...(21)


sk
2
ln (1 + 1t + 2 t + 3 3
t + ...) =  tk
k 1 k

1 1
 
= 1t (12  22 )t2  ( 1
3
 3...(22)
2 3
so the first few values are ...(23)

s= 
...(24)
1 1

s2 = 12  22 ...(25)

s3 = 13 – 312  33

4 2 2
s4 = 1  412  22  413  44 .
In general, s can be computed from the determinant
p

1 1

22 1

s = (1) p1 33 2 ...(26)

p

44 3


pp p1

LOVELY PROFESSIONAL UNIVERSITY

95
 Unit 13 : The Galois Group of a Polynomial

(Littlewood 1958, Cadogan 1971). In particular, Notes

n
S ...(27)
1 (x1,..., xn) = xk  1

k1
...(28)
S (x ,..., x ) =  2  2
2 1 n 1 2 ...(29)
S (x ,..., x ) =  3  3   3
...(30)
3 1 n 1 1 2 3

S (x ,..., x ) =  4  42  2 2

4 1 n 1 1 2 2

(Schroeppel 1972), as can be verified by plugging in and multiplying

through.
13.4 Constructible Polygon

In mathematics, a constructible polygon is a regular polygon that can be

constructed with compass and straightedge. For example, a regular pentagon is
constructible with compass and straightedge while a regular heptagon is not.

Conditions for Constructibility

Some regular polygons are easy to construct with compass and straightedge;
others are not. This led to the question being posed: is it possible to construct all
regular n-gons with compass and straightedge? If not, which n-gons are
constructible and which are not?
Carl Friedrich Gauss proved the constructability of the regular 17-gon in 1796. Five
years later, he developed the theory of Gaussian periods in his Disquisitiones
Arithmeticae. This theory allowed him to formulate a sufficient condition for the
constructability of regular polygons.
A regular n-gon can be constructed with compass and straight edge if n is the
product of a power of 2 and any number of distinct Fermat primes.
Gauss stated without proof that this condition was also necessary, but never published his proof.
A full proof of necessity was given by Pierre Wantzel in 1837. The result is known as the Gauss–
Wantzel theorem.
Figure 3.1

LOVELY PROFESSIONAL UNIVERSITY

96
Notes Construction of the regular 257-gon
Detailed results by Gauss’ theory
Only five Fermat primes are known:
F = 3, F = 5, F = 17, F = 257, and F = 65537 (sequence A019434 in OEIS)
0 1 2 3 4

The next twenty-eight Fermat numbers, F through F , are known to be composite.

5 32

Thus an n-gon is constructible if

n = 3, 4, 5, 6, 8, 10, 12, 15, 16, 17, 20, 24, … (sequence A003401 in OEIS),
while an n-gon is not constructible with compass and straightedge if
n = 7, 9, 11, 13, 14, 18, 19, 21, 22, 23, 25, … (sequence A004169 in OEIS).

13.5 Connection to Pascal’s Triangle

There are 31 known numbers that are multiples of distinct Fermat primes, which
correspond to the 31 odd-sided regular polygons that are known to be constructible.
These are 3, 5, 15, 17, 51, 85, 255, 257, …, 4294967295. As John Conway commented in
The Book of Numbers, these numbers, when written in binary, are equal to the first 32
rows of the modulo-2 Pascal’s triangle, minus the top row. This pattern breaks down
after there, as the 6th Fermat number is composite, so the following rows do not
correspond to constructible polygons. It is unknown whether any more Fermat primes
exist, and is therefore unknown how many odd-sided constructible polygons exist. In
general, if there are x Fermat primes, then there are 2 x”1 odd-sided constructible
polygons.

General Theory

In the light of later work on Galois Theory, the principles of these proofs have been
clarified. It is straightforward to show from analytic geometry that constructible
lengths must come from base lengths by the solution of some sequence of
quadratic equations. In terms of field theory, such lengths must be contained in a
field extension generated by a tower of quadratic extensions. It follows that a field
generated by constructions will always have degree over the base field that is a
power of two.
In the specific case of a regular n-gon, the question reduces to the question of
constructing a length
cos(2/n).
This number lies in the n-th cyclotomic field — and in fact in its real subfield, which
is a totally real field and a rational vector space of dimension
½(n),
where (n) is Euler’s quotient function. Wantzel’s result comes down to a
calculation showing that (n) is a power of 2 precisely in the cases specified.
As for the construction of Gauss, when the Galois group is 2-group it follows that it
has a sequence of subgroups of orders
1, 2, 4, 8, ...
that are nested, each in the next something simple to prove by induction in this case of
an abelian group. Therefore, there are subfields nested inside the cyclotomic field, each
of degree 2 over the one before. Generators for each such field can be written down by
Gaussian period theory.

LOVELY PROFESSIONAL UNIVERSITY

97
 Unit 13: The Galois Group of a
Polynomial

For example for n = 17 there is a period that is a sum of eight roots of unity, one that is a sum of Notes
four roots of unity, and one that is the sum of two, which is cos(2/17).
Each of those is a root of a quadratic equation in terms of the one before. Moreover, these
equations have real rather than imaginary roots, so in principle can be solved by geometric
construction: this because the work all goes on inside a totally real field.
In this way the result of Gauss can be understood in current terms; for actual calculation of the
equations to be solved, the periods can be squared and compared with the ‘lower’ periods, in a
quite feasible algorithm.

Compass and Straightedge Constructions

Compass and straightedge constructions are known for all constructible polygons. If
n = p· q with p = 2 or p and q co-prime, an n-gon can be constructed from a p-gon
and a q-gon.
 If p = 2, draw a q-gon and bisect one of its central angles. From this, a 2q-gon can be constructed.
 If p > 2, inscribe a p-gon and a q-gon in the same circle in such a way that they share a vertex. Because p and q
are relatively prime, there exists integers a,b such that ap + bq = 1. Then 2a/q + 2b/p = 2/pq. From this, a p·q-
gon can be constructed.
Thus one only has to find a compass and straightedge construction for n-gons
where n is a Fermat prime.
 The construction for an equilateral triangle is simple and has been known since Antiquity. Constructions for the regular
pentagon were described both by Euclid and by Ptolemy.
 Although Gauss proved that the regular 17-gon is constructible, he didn’t actually show how to do it. The first
construction is due to Erchinger, a few years after Gauss’ work.
 The first explicit construction of a regular 257-gon was given by Friedrich Julius Richelot (1832).
 A construction for a regular 65537-gon was first given by Johann Gustav Hermes (1894). The construction is
very complex; Hermes spent 10 years completing the 200-page manuscript. (Conway has cast doubt on the validity
of Hermes’ construction, however.

Figure 13 .2

LOVELY PROFESSIONAL UNIVERSITY

98
Notes
Figure 13 .3

From left to right, constructions of a 17-gon, 257-gon and 65537-gon.

Other Constructions

It should be stressed that the concept of constructible as discussed in this article applies
specifically
to compass and straightedge construction. More constructions become possible if other tools
are allowed. The so-called neusis constructions, for example, make use of a marked
rulers.The
constructions are a mathematical idealization and are assumed to be done exactly.

Example: Determine the group of all automorphisms of a field with 4 elements.

Solution: The automorphism group consists of two elements: the identity mapping
and the Frobenius automorphism.
As you know this field with 4 elements can be constructed as F = Z [x] / < x 2+x+1 >. Letting a
be
2

the coset of x, we have F = {0, 1, a, 1+a}. Any automorphism of F must leave 0

and 1 fixed, so the only possibility for an automorphism other than the identity is to
interchange a and 1+a. Is this an automorphism? Since x 2+x+1 0, we have x2 -x-1
x+1, so a2 = 1+a and (1+a)2 = 1+2a+a2 = a. Thus the function that fixes 0 and 1
while interchanging a and 1+a is in fact the Frobenius automorphism of F.

Example: Let F be the splitting field in C of x4+1.

(i) Show that [F:Q] = 4.
Solution: The polynomial x8-1 factors over Q as x8-1 = (x4-1)(x4+1) = (x-1)(x+1)(x2+1)
(x4 +1). The factor x4 +1 is irreducible over Q by Eisenstein’s criterion. The roots of x 4+1
are thus the primitive
8th roots of unity, ± 2 / 2 ± 2 / 2i, and adjoining one of these roots also gives the
others, together with i. Thus, the splitting field is obtained in one step, by adjoining one
root of x4+1, so its degree over Q is 4.

It is clear that the splitting field can also be obtained by adjoining first 2 and then i,
so it can also be expressed as Q( 2 , i).

LOVELY PROFESSIONAL UNIVERSITY

99
 Unit 13 : The Galois Group of a Polynomial

Notes
(ii) Find automorphisms of F that have fixed fields Q( 2 ), Q(i), and Q( 2 i), respectively.

Solution: These subfields of Q( 2 , i) are the splitting fields of x2-2, x2+1, and x2+2, respectively.
Any automorphism must take roots to roots, so if is an automorphism of Q( 2 , i), we must
have ( 2 ) = ± 2 , and (i) = ± i. These possibilities must in fact define 4
automorphisms of the splitting field.

If we define 1 ( 2 ) = 2 and  (i) = -i, then the subfield fixed by  is Q( 2 ). If we define

1 1

( 2)=– 2 and  (i) = i, then the subfield fixed by  is Q(i). Finally, for  =   we have
2 2 2 3 2 1

( 2)=– 2 and (i) = –i, and thus  ( 2 i) = 2 i, so  has Q( 2 i) as its fixed subfield.
3 3 3

Example: Find the Galois groups of x3 – 2 over the fields Z5 and Z11.
3 2
Solution: The polynomial is not irreducible over Z , since it factors as x -2 = (x+2)(x -2x-1). The
5
quadratic factor will have a splitting field of degree 2 over Z , so the Galois group is cyclic of
5
order 2.
A search in Z for roots of x3-2 yields one and only one: x = 7. Then x3-2 can be factored as x3-2 =
11
(x-7)(x2+7x+5), and the second factor must be irreducible. The splitting field has degree 2 over
Z , and can be described as Z [x] / < x2+7x+5 >. Thus the Galois group is cyclic of order 2.
11 11

Example: Find the Galois group of x4-1 over the field Z .

7

Solution: We first need to find the splitting field of x4-1 over Z . We have x4-1 = (x-1)(x+1)(x2+1).
7
A quick check of ±2 and ±3 shows that they are not roots of x 2+1 over Z , so x2+1 is irreducible
7
over Z . To obtain the splitting field we must adjoin a root of x 2+1, so we get a splitting field
7
2
Z [x] / < x +1 > of degree 2 over Z .
7 7

The Galois group of x4-1 over Z is cyclic of order 2.

7

Example: Find the Galois group of x3-2 over the field Z .

7

Solution: In this case, x3-2 has no roots in Z , so it is irreducible. We first adjoin a root a of x 3-2 to
7
Z . The resulting extension Z (a) has degree 3 over Z , so it has 73 = 343 elements, and each
7 7 7
element is a root of the polynomial x343-x. Let b> be a generator of the multiplicative group of
114 3 342
the extension. Then (b ) =b = 1, showing that Z (a) contains a non-trivial cube root of 1. It
7
follows that x3-2 has three distinct roots in Z (a): a, ab114, and ab228, so therefore Z (a) is a splitting
7 7
field for x3-2 over Z . Since the splitting field has degree 3 over Z , it follows the Galois group of
7 7
the polynomial is cyclic of order 3.

Self Assessment

1. Galois considered ................... of the roots that leave the coefficient field fixed.
(a) polynomial (b) permutation
(c) combination (d) range
2. The modern approach is to consider ................... determined by permutation.
(a) homomorphism (b) automorphism
(c) isomorphism (d) ideal and subfield

LOVELY PROFESSIONAL UNIVERSITY

100
 Notes 3. Any automorphism of a field f must leave its prime ................... fixed.
(a) sub group (b) sub domain
(c) sub field (d) sub range
4. Given [F : Q] is equal to ...................

(a) 7 (b)
(c) 6 (d)
5. Automorphism of F that have fixed fields Q( 2),Q(i) and Q 2,i respectively.

(a) Q1  Q 2,Q2  Q(i),Q3 2i  2i

(b) Q1  2,Q2  i,Q3 2i  0
(c) Q=Q =Q
2 3

(d) Q =Q–1=Q–1
1 2 3

13.6 Summary

 Let F be an extension field of K. The set of all automorphisms  : F > F such that (a)
= a for all a in K is a group under composition of functions.
 Let F be an extension field of K. The set
{ in Aut(F) | (a) = a for all a in K }
is called the Galois group of F over K, denoted by Gal(F/K).
 Let K be a field, let f(x) be a polynomial in K[x], and let F be a splitting field for f(x)
over
K. Then Gal(F/K) is called the Galois group of f(x) over K, or the Galois group of
the equation f(x) = 0 over K.
 Let F be an extension field of K, and let f(x) be a polynomial in K[x]. Then any
element of Gal(F/K) defines a permutation of the roots of f(x) that lie in F.
 Let f(x) be a polynomial in K[x] with no repeated roots and let F be a splitting field for
f(x) over K. If  : K > L is a field isomorphism that maps f(x) to g(x) in L[x] and E is a splitting field for g(x) over
L, then there exist exactly [F:K] isomorphisms : F -> E such that (a) = (a) for all a in K.
 Let K be a field, let f(x) be a polynomial in K[x], and let F be a splitting field for f(x)
over
K. If f(x) has no repeated roots, then |Gal(F/K)| = [F:K].
 Let K be a finite field and let F be an extension of K with [F:K] = m. Then Gal(F/K) is a
cyclic group of order m.
 If we take K = Z , where p is a prime number, and F is an extension of degree m,
then the
p
generator of the cyclic group Gal(F/K) is the automorphism  : F -> F defined
by (x) = xp, for all x in F. This automorphism is called the Frobenius
automorphism of F.

13.7 Keywords

Galois Group: Let F be an extension field of K. The set

{ in Aut(F) | (a) = a for all a in K }
is called the Galois group of F over K, denoted by Gal(F/K).

LOVELY PROFESSIONAL UNIVERSITY

101
 Unit 13 : The Galois Group of a Polynomial

Galois Group of the Equation: Let K be a field, let f(x) be a polynomial in K[x], and let F be a Notes
splitting field for f(x) over K. Then Gal(F/K) is called the Galois group of f(x) over K, or the
Galois group of the equation f(x) = 0 over K.

13.8 Review Questions

1. Let p be prime. Prove that there exists a polynomial f(x)  Q[x] of degree p with Galois

group isomorphic to S . Conclude that for each prime p with p  5 there exists a polynomial
p
of degree p that is not solvable by radicals.
2. Let p be a prime and Z (t) be the field of rational functions over Z . Prove that f(x) = x p – t
p p
is an irreducible polynomial in Zp(t)[x]. Show that f(x) is not separable.
3. Let E be an extension field of F. Suppose that K and L are two intermediate fields. If there exists an element  
G(E/F) such that (K) = L, then K and L are said to be conjugate fields. Prove that K and L are conjugate if and only if
G(E/K) and G(E/L) are conjugate subgroups of G(E/F).
4. Let   Aut( ). If a is a positive real number, show that (a) > 0.
5. Let K be the splitting field of x3 + x2 + 1  [x]. Prove or disprove that K is an extension by
2
radicals.
6. Let F be a field such that char F  2. Prove that the splitting field of f(x) = ax2 + bx + c is F( ), where a = b2 –
4ac.
7. Prove or disprove: Two different subgroups of a Galois group will have different fixed fields.
8. Let K be the splitting field of a polynomial over F. If E is a field extension of F contained in K and [E : F] = 2, then
E is the splitting field of some polynomial in F[x].

Answers: Self Assessment

1. (b) 2. (b) 3. (c)4. (d) 5. (a)

13.9 Further Readings

Books
Dan Saracino: Abstract Algebra; A First Course.
Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links
www.jmilne.org/math/CourseNotes/
www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

102
 Richa Nandra, Lovely Professional University

Notes Unit 14: Solvability by Radicals

CONTENTS
Objectives
Introduction
14.1 Radical Extension
14.2 Solvable by Radicals
14.3 Summary
14.4 Keywords
14.5 Review Questions
14.6 Further Readings

Objectives

After studying this unit, you will be able to:

 Discuss the radical extension
 Explain that a polynomial equation is solvable by radical

Introduction

In most results, in this section we will assume that the fields have characteristic
zero, in order to guarantee that no irreducible polynomial has multiple roots. When
we say that a polynomial equation is solvable by radicals, we mean that the
solutions can be obtained from the coefficients in a finite sequence of steps, each of
which may involve addition, subtraction, multiplication, division, or taking nth roots.
Only the extraction of an nth root leads to a larger field, and so our formal definition
is phrased in terms of subfields and adjunction of roots of x n-a for suitable elements
a.

14.1 Radical Extension

Definition: An extension field F of K is called a radical extension of K if there

exist elements u , u , ... , u in F and positive integers n , n , ... , n such that
1 2 m 1 2 m

(i) F = K (u , u , ... , u ), and

1 2 m
n
(ii) un is in K and u is in K ( u , ... , u ) for i = 2, ... , m
11 ii 1 i-1

14.2 Solvable by Radicals

For a polynomial f(x) in K[x], the polynomial equation f(x) = 0 is said to be solvable
by radicals if there exists a radical extension F of K that contains all roots of f(x).
Proposition: Let F be the splitting field of xn - 1 over a field K of characteristic
zero. Then Gal(F/K) is an abelian group.

LOVELY PROFESSIONAL UNIVERSITY

103
 Unit 14: Solvability by Radicals

Theorem 1: Let K be a field of characteristic zero that contains all nth roots of unity, let a be an Notes
element of K, and let F be the splitting field of xn-a over K. Then Gal(F/K) is a cyclic group whose
order is a divisor of n.
Theorem 2: Let p be a prime number, let K be a field that contains all pth roots of unity, and let
F be an extension of K. If [F:K] = |Gal(F/K)| = p, then F = K(u) for some u in F such that u p is in
K.
Lemma: Let K be a field of characteristic zero, and let E be a radical extension of K. Then there
exists an extension F of E that is a normal radical extension of K.
Theorem 3: Let f(x) be a polynomial over a field K of characteristic zero. The equation
f(x) = 0 is solvable by radicals if and only if the Galois group of f(x) over K is solvable.
S is not solvable for n  5, and so to give an example of a polynomial equation of degree n that
n

is not solvable by radicals, we only need to find a polynomial of degree n whose Galois group
over Q is S .
n

Lemma: Any subgroup of S that contains both a transposition and a cycle of length 5 must be
5

equal to S itself.
5

Theorem 4: There exists a polynomial of degree 5 with rational coefficients that is not solvable
by radicals

Example: Let f(x) be irreducible over Q, and let F be its splitting field over Q. Show that if
Gal (F/Q) is abelian, then F = Q(u) for all roots u of f(x).
Solution: Since F has characteristic zero, we are in the situation of the fundamental theorem of
Galois theory. Because Gal (F/Q) is abelian, every intermediate extension between Q and F must
be normal. Therefore, if we adjoin any root u of f(x), the extension Q(u) must contain all other
roots of f(x), since it is irreducible over Q. Thus Q(u) is a splitting field for f(x), so Q(u) = F.

Example: Find the Galois group of x9-1 over Q.

Solution: We can construct the splitting field F of x9-1 over Q by adjoining a primitive 9th root
of unity to Q. We have the factorization

x9-1 = (x3-1)(x6+x3+1)
= (x-1)(x2+x+1)(x6+x3+1).
Substituting x+1 in the last factor yields

(x+1)6+(x+1)3+1 = x6+6x5+15x4+ 21x3+18x2+9x+3.

This polynomial satisfies Eisenstein’s criterion for the prime 3, which implies that the factor
x6+x3+1 is irreducible over Q. The roots of this factor are the primitive 9th roots of unity, so it
× ×
follows that [F:Q] = 6. Gal (F/Q) is isomorphic to a subgroup of Z Since Z is abelian of order
9 9
6, it is isomorphic to Z . It follows that Gal (F/Q) Z.
6 6

Comment: The Galois group of xn-1 over Q is isomorphic to Z ×

and so the Galois group is cyclic
n
k k
of order (n) iff n = 2, 4, p , or 2p , for an odd prime p.

Example: Show that x4-x3+x2-x+1 is irreducible over Q, and use it to find the
Galois group of x10-1 over Q.
LOVELY PROFESSIONAL UNIVERSITY

104
 Solution: We can construct the splitting field F of x10-1 over Q by adjoining a primitive 10th
Notes root
of unity to Q. We have the factorization

x10-1 = (x5-1)(x5+1)
= (x-1)(x4+x3+x2+x+1) (x+1)(x4-x3+x2-x+1).
Substituting x-1 in the last factor yields

(x-1)4-(x-1)3+(x-1)2-(x-1)+1
= (x4-4x3+6x2-4x+1) - (x3-3x2+3x-1) + (x2-2x+1) - (x-1) + 1
= x4-5x3+10x2-10x+5.
This polynomial satisfies Eisenstein’s criterion for the prime 5, which implies that the factor
x4-x3+x2-x+1 is irreducible over Q.
The roots of this factor are the primitive 10th roots of unity, so it follows that [F:Q] = 4. The
proof
of Theorem 1 shows that Gal (F/Q) Z10× and so the Galois group is cyclic of order 4.

Example: Show that p(x) = x5-4x+2 is irreducible over Q, and find the number of rea
roots. Find the Galois group of p(x) over Q, and explain why the group is not solvable.
Solution: The polynomial p(x) is irreducible over Q since it satisfies Eisenstein’s criterion for
p = 2. Since p(-2) = -22, p(-1) = 5, p(0) = 2, p(1) = –1, and p(2) = 26, we see that p(x) has a
real root
between -2 and -1, another between 0 and 1, and a third between 1 and 2. The derivative
p’(x) = 5x4-4 has two real roots, so p(x) has one relative maximum and one relative minimum
and thus it must have exactly three real roots. It follows as in the proof of Theorem 2 that the
Galois group of p(x) over Q is S , and so it is not solvable.
5

Self Assessment

1. Let F be splitting field of xn – 1 over a field K of characteristic .................. then G(F/K)

is an abelian group.
(a) 1 (b)
(c) –1 (d)
2. Let K be a field of characteristic zero and let  be a .................. of K. Thus there
exists an extension of F of  that is normal radical extension.
(a) radical extension (b) solvable group
(c) Galois group (d) finite element
3. There exists a polynomial of degree .................. with rational co-efficients that is not
solvable by radical.
(a) 4 (b)
(c) 6 (d)
4. Any subgroup of S that contains both a transposition and cycle of length ..................
must
5

be equal to S itself.
5

(a) 4 (b)
(c) 3 (d)
LOVELY PROFESSIONAL UNIVERSITY

105
 Unit 14: Solvability by Radicals

Notes
14.3 Summary
 An extension field F of K is called a radical extension of K if there exist elements
u , u , ... , u in F and positive integers n , n , ... , n such that
1 2 m 1 2 m

(i) F = K (u , u , ... , u ), and

1 2 m
n n
(ii) u is in K and u is in K ( u , ... , u ) for i = 2, ... , m .
11 ii 1 i-1

For a polynomial f(x) in K[x], the polynomial equation f(x) = 0 is said to be

solvable by radicals if there exists a radical extension F of K that contains
all roots of f(x).
 Let F be the splitting field of xn - 1 over a field K of characteristic zero. Then Gal(F/K) is an abelian group.
 Let K be a field of characteristic zero that contains all nth roots of unity, let a be an element of K, and let F be
the splitting field of xn-a over K. Then Gal(F/K) is a cyclic group whose order is a divisor of n.
 Let p be a prime number, let K be a field that contains all pth roots of unity, and let F be an extension of K. If [F:K] = |
Gal(F/K)| = p, then F = K(u) for some u in F such that u p is in K.
 Let K be a field of characteristic zero, and let E be a radical extension of K. Then there exists an extension F of E
that is a normal radical extension of K.
 Let f(x) be a polynomial over a field K of characteristic zero. The equation f(x) = 0 is solvable by radicals if and
only if the Galois group of f(x) over K is solvable.
 S is not solvable for n  5, and so to give an example of a polynomial equation of degree
n
n that is not solvable by radicals, we only need to find a polynomial of degree
n whose Galois group over Q is S .
n

 Any subgroup of S that contains both a transposition and a cycle of length 5 must be equal
5
to S itself.
5

 There exists a polynomial of degree 5 with rational coefficients that is not solvable by radicals

14.4 Keywords

Radical Extension: An extension field F of K is called a radical extension of K if

there exist elements u , u , ... , u in F and positive integers n , n , ... , n such that
1 2 m 1 2 m

(i) F = K (u , u , ... , u )
1 2 m

Solvable by Radicals: For a polynomial f(x) in K[x], the polynomial equation f(x) =
0 is said to be solvable by radicals if there exists a radical extension F of K that
contains all roots of f(x).

14.5 Review Questions

1. We know that the cyclotomic polynomial

p
x 1 p1 p2

p (x)  x x  ...  x  1

is irreducible over for every prime p. Let w be a zero  (x), and consider the field ().
p

(a) Show that , 2,...,p-1 are distinct zeros of  (x), and conclude that they are all the
p
zeros of  (x).
p

LOVELY PROFESSIONAL UNIVERSITY

106
 Notes (b) Show that G( (w)/ ) is abelian of order p – 1.
(c) Show that the fixed field of G( ()/ ) is .
2. Let F be a finite field of characteristic zero. Let E be a finite normal extension of F
with Galois group G(E/F): Prove that F  K  L  E if and only if {id}  G(E/L)  G(E/K)  G(E/F).
3. Let F be a field of characteristic zero and let f(x)  F[x] be a separable polynomial of
degree
n. If E is the splitting field of f(x), let  ,..., be the roots of f(x) in E. Let   
1 n
We define the discriminant of f(x) to be 2.
(a) If f(x) = ax2 + bx + c, show that 2 = b2 – 4ac.
(b) If f(x) = x3 + px + q, show that 2 = –4p3 – 27q2.
(c) Prove that 2 is in F.
(d) If   G(E/F) is a transposition of two roots of f(x), show that () = –.
(e) If   G(E/F) is an even permutation of the roots of f(x), show that () = .
(f) Prove that G(E/F) is isomorphic to a subgroup of A if and only if   F.
n

(g) Determine the Galois groups of x + 2x – 4 and x3 + x – 3.

3

Answers: Self Assessment

1. (b) 2. (a) 3. (b) 4. (b)

14.6 Further Readings

Books Dan Saracino: Abstract Algebra; A First Course.

Mitchell and Mitchell: An Introduction to Abstract Algebra.
John B. Fraleigh: An Introduction to Abstract Algebra (Relevant Portion).

Online links www.jmilne.org/math/CourseNotes/

www.math.niu.edu
www.maths.tcd.ie/
archives.math.utk.edu

LOVELY PROFESSIONAL UNIVERSITY

107