Notes in ring theory

Paul Martin
Dec 11, 2009 (printed: May 8, 2012)

Contents
1 Foreword

2 Rings, Polynomials and Fields

2.1 Introduction . . . . . . . . . . . . . . . . . . . .
2.1.1 Some remarks on sets and relations . . .
2.2 Factorisation of integers . . . . . . . . . . . . .
2.3 Rings . . . . . . . . . . . . . . . . . . . . . . .
2.3.1 Ring homomorphisms . . . . . . . . . .
2.4 Factorisation in integral domains . . . . . . . .
2.5 Polynomials . . . . . . . . . . . . . . . . . . . .
2.6 Polynomials over Z . . . . . . . . . . . . . . . .
2.7 Irreducible polynomials . . . . . . . . . . . . .
2.8 Fields of fractions . . . . . . . . . . . . . . . . .
2.9 Extension Fields . . . . . . . . . . . . . . . . .
2.9.1 Extending a field by algebraic elements
2.9.2 Remarks on Kronecker . . . . . . . . . .
2.9.3 Geometric constructions . . . . . . . . .
2.10 Ideals . . . . . . . . . . . . . . . . . . . . . . .
2.11 More Exercises . . . . . . . . . . . . . . . . . .
2.12 More revision exercises . . . . . . . . . . . . . .
2.13 Homework exercises . . . . . . . . . . . . . . .
2.14 More (Optional) Rings and Exercises . . . . . .
2.14.1 Group rings and skew group rings . . .
2.14.2 Monoid-graded algebras . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

7
7
8
10
11
12
16
19
20
21
22
23
23
25
26
26
30
30
31
34
35
35

CONTENTS

Chapter 1

Foreword
Ring theory is generally perceived as a subject in Pure Mathematics. This means that it is a
subject of intrinsic beauty. However, the idea of a ring is so fundamental that it is also vital
in many applications of Mathematics. Indeed it is so fundamental that very many other vital
tools of Applied Mathematics are built from it. For example, the crucial notion of linearity, and
linear algebra, which is a practical necessity in Physics, Chemistry, Biology, Finance, Economics,
Engineering and so on, is built on the notion of a vector space, which is a special kind of ring
module.
At Undergraduate Level Three and beyond, one typically encounters many applications of ring
theory (either explicitly or implicitly). For example, many fundamental notions about information
and information transmission (not to mention information protection) are most naturally described
in the setting of ring theory. In particular, a field is a special kind of ring, and the theory of Coding
one of the main planks of modern information technology and Computer Science makes
heavy practical use of the theory of fields, which lives inside the theory of rings.
So, there are countless applications of ring theory ahead (not to mention countless amazing
open problems). But here we shall concentrate, for now, on the first point: ring theory is a
subject of intrinsic beauty.
Ring theory appears to have been among the favourite subjects of some of the most influential
Scientists of the twentieth century, such as Emmy Noether (discoverer both of Noethers Theorem
one of the most important theorems in modern Physics; and of Noetherian rings1 ); and Alfred
Goldie (author of Goldies Theorem, and founder of the University of Leeds Algebra Group).
But perhaps more important than any of these points is that ring theory is a core part of
the subject of Algebra, which forms the language within which modern Science can be put on its
firmest possible footing.

1 Also

someone who helped to defeat the terrible sexism that afflicted European academia in the 20th century.

CHAPTER 1. FOREWORD

Chapter 2

Rings, Polynomials and Fields

This Chapter is based partly on the undergraduate lecture course notes of Bill Crawley-Boevey, and
sections from the textbooks of Serge Lang and Nathan Jacobson. It is intended as an undergraduate
exposition.

2.1

Introduction

As we shall see later, a ring is a set with two binary operations (usually called addition and
multiplication) satisfying certain axioms. The most basic example is the set Z of integers. Another
familiar example is the set Z[x] of polynomials in an indeterminate x, with integer coefficients.
(2.1.1) Example. Make sure you can add and multiply polynomials, by trying a few examples.

(2.1.2) Example. Suppose n Z is not a square and define Z[ n] = {a + b n | a, b Z}. Check

that you can add and multiply in this case, and that these operations are closed.
(2.1.3) Let us start by thinking about the integers before we go any further. They are (hopefully)
familiar, and familiarity breeds contempt.1 But what aspects of their behaviour should we not
be contemptuous of? What we need to do, for a moment, is forget the familiarity, and consider
the integers as an example of an algebraic structure. What can we say about them in this light?
Addition and multiplication mean that, given a, b Z we have solutions x, y Z to the
equations
x = a + b,
y = ab
Continuing to regard a, b as given, this does not automatically mean that we have a solution x to
xa = b

(2.1)

It also does not automatically mean that we have a solution x to x + a = b, but of course if a, b Z
then this problem does have a solution. On the other hand equation (2.1) does not always have a
solution. It depends on a, b. For this reason, we say a divides b if xa = b has a solution x Z. If
a divides b we may write this as a|b.
1 According

to Charles Dickens.

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

12

16

24

10

14

18

20 27 28 30

15

21

11

13

22

25 26

17

...

Figure 2.1: The start of the divides Hasse diagram

(2.1.4) Recall that a poset is a set S together with a transitive, reflexive and antisymmetric relation
on S. If is such a relation we write a > b if a b and a 6= b.
A transitive reduction of a poset (S, ) is a simple directed graph with vertex set S and an
edge (a, b) if a > b and there does not exist a > c > b.
A lattice is a poset such that for any pair of elements (a, b) there is a least upper bound (LUB)
and a greatest lower bound (GLB).
(2.1.5) We define a relation on N by a b if a|b. This is evidently transitive, reflexive and
antisymmetric. Thus it makes N into a poset (different from the obvious but very important total
order (N, )). Indeed (N, |) is a (complete distributive) lattice. In (N, |) the GLB is the GCD.

(2.1.6) Example. The GCD of 6 and 8 is 2. A lower bound of 6 in this case is a number that
divides 6 (thus 6,3,2 or 1). A lower bound of 8 is any of 8,4,2,1. Thus a lower bound of {6, 8} is
an element of {1, 2}. Since 1|2 but 2 6 |1 we have the GCD of 2.

(2.1.7) We say that a, b are coprime if their GCD is 1.

(2.1.8) Exercise. Draw the begining of the Hasse diagram for (N, |). (See Figure 2.1.)

2.1.1

Some remarks on sets and relations

(This Section contains remarks and reminders only. It can safely be skipped if you are in a hurry
to get on with the ring theory.)
Suppose we have an arbitrary poset (S, ). What does the term lower bound mean?
(2.1.9) By convention if a relation (S, ) is a poset then ab can be read as: a is less than or
equal to b. This emphasises that the relation has a direction, even if the symbol itself does not.
Recalling the inverse (or dual) relation 1 , one could write a1 b for ba.

2.1. INTRODUCTION

Some relation symbols come with their own inverse symbol. For example by convention one
could consider a b as another way of writing b a in the poset (S, ), even if does not have
one of its usual meanings (see e.g. Howie 2 ).
The less than convention allows us to make sense of the term lower bound, as in: c is a lower
bound of {a, b} if ca and cb. However the inverse relation to a poset is also a poset, and one
sees that inversion swaps the roles of lower and upper bounds. In this sense there is a symmetry
between them, as concepts. If a relation has both, then its inverse also has both (although it is
not necessarily an isomorphic relation).
Altogether, though, this means that we should be a little careful with notation. If we declare
a poset (S, ), do we really want to read a b as a is less than b ? Or do we really want to be
careful about our intrinsic reading of the symbol , and understand (S, ) as the intended poset
to which the convention applies?
Abstractly we are free to choose. But then we should spell out which is upper and which is
lower bound (and our choice might sometimes be counter-intuitive). Our definition of lattice above
has both LUB and GLB, so it works either way. (But for definiteness let us say that we chose the
second alternative!)
Here are some definitions and facts from basic set theory.
(2.1.10) The power set P (S) of a set S is the set of all subsets.
(2.1.11) A relation on S is an element of P (S S).
(2.1.12) A preorder is a reflexive transitive relation. Thus a poset is an antisymmetric preorder;
and an equivalence is a symmetric preorder.
An ordered set is a poset with every pair of elements comparable. A well-ordered set is an
ordered set such that every subset has a least element.
Example: (R, ) is ordered but not well-ordered; (N, ) is well-ordered but the opposite relation
is only ordered. (Z, ) is ordered but not well-ordered; Z ordered as 0, 1, 1, 2, 2, 3, ... is wellordered; Z ordered as 0, 1, 2, 3, ..., 1, 2, 3, ... is well-ordered.
Can R be well-ordered (with a different order)? Good question! In fact the answer requires a
more careful formulation of set theory than we have assumed, so we will have to leave it. (This
will not be a problem for us in studying basic ring theory, but problems of Algebra often do drive
the study of problems in Logic!)
(2.1.13) Suppose S, T are preorders. A map f : S T is order preserving if x y in S implies
f (x) f (y) in T . (Note our challenging use of notation here.)
(2.1.14) An interval in an arbitrary poset (S, ) is defined analogously to the case (R, ). That
is, [x, y] := {z S | x z y}. The open interval (x, y) is defined similarly. See e.g. [?].
A poset is locally finite if every interval is finite.
(2.1.15) The poset (P (S), ) is a lattice.
(2.1.16) A topology on a set S is subset T of P (S) such that (i) S and are in T (ii) A finite
intersection of sets in T is in T (iii) A union of sets in T is in T .
The open sets of S in the topology T are the elements of T . A subset of S is closed if it is the
complement of an open set. A topological space is a pair (S, T ) as above.
2J

M Howie, Fundamentals of semigroup theory, Oxford 1995.

10

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.1.17) An Alexandroff topology is a topology that is closed under intersection. (Thus a finite
topology is necessarily Alexandroff.)
Such a topology on S defines a relation on S: Let x be the intersection of open sets containing
x. Then x y if x y (that is, if x y ).
(2.1.18) A T0 topology is a topology (S, T ) such that for any pair s, t S there is an X T
containing precisely one of s, t.
(2.1.19) For any set S the trivial topology is T = {, S}; and the discrete topology is T = P (S).

(2.1.20) Example. The trivial topology is not T0 unless |S| = 1.

The topologies on S = {0, 1} are {, S}; {, S, {0}}; {, S, {1}}; {, S, {0}, {1}}. All apart from
the trivial topology are T0 .
(2.1.21) A function f : S S between the underlying sets in two topologies is continuous if J
an open set in S implies f 1 (J) open in S.
(2.1.22) Lemma. The relation on S defined in (2.1.17) above is a preorder. A function f : S S
between the underlying sets in two Alexandroff topologies is continuous iff it is order preserving
on the induced preorder.

2.2

Factorisation of integers

(2.2.1) A prime number is an integer p > 1 such that p|ab implies p|a or p|b.
(2.2.2) Theorem. (Fundamental theorem of arithmetic) Any integer n > 1 can be written as a
product of prime numbers:
n = p1 p2 ...pk
And if n = q1 q2 ...ql is another product of primes then k = l and indeed the factorisations are the
same up to reordering.
Proof. First we show that such a factorisation is always possible (then later we will show the
uniqueness up to reordering).
Suppose, for a contradiction, that there is an n that cannot be factorised into prime numbers.
Then there is a least such n (note that we are using an ordering property of the integers here that
goes beyond just arithmetic in fact we have already used the > ordering repeatedly!). Of course
any such n cannot itself be prime (else n = n would already be an acceptable prime factorisation).
Since it is not prime, it has a divisor d (say), that is d|n and 1 < d < n. So n = dm, with
1 < m < n. But since n was the least integer without a factorisation, both d and m have one. But
then the combination of these is a factorisation of n a contradiction.
To show uniqueness, we work again for a contradiction. Let n be the lowest integer with distinct
factorisations, and let them be n = p1 ...pk and n = q1 ...ql . If k = 1 then n is prime, so l = 1 and
p1 = q1 giving a contradiction, so k > 1. But in this case note that p1 |q1 ...ql . Thus by definition
p1 |qi for some i. That is, p1 = qi and n = mp1 = mqi for some m. Cancelling p1 from each
factorisation we get two factorisations for m. But 1 < m < n, so m has a unique factorisation
a contradiction. 2
(2.2.3) Theorem. (Euclid) There are infinitely many prime numbers.

2.3. RINGS

11

Proof. Exercise. Hint: Suppose p1 p2 ...pn is a product of primes. Then p1 p2 ...pn + 1 is coprime to
them all. Now suppose, for a contradiction, that there are only finitely many primes.
(2.2.4) Perhaps it is strange that the integers contain this special prime structure, whereas the
larger sets such as the rationals Q and the reals R do not (it is too easy to solve xa = b in these
cases).

2.3

Rings

(2.3.1) A monoid is a set with a closed associative binary operation, with an identity element.
(2.3.2) Example. (Z, ) is a monoid so this is an even simpler thing than a ring, because we
only use one binary operation.
(2.3.3) A group is a monoid (G, , 1) with the property that every element g G has an inverse
(i.e. there is an element g G such that g g = 1).
A group is abelian if a b = b a.
(2.3.4) A ring is a set R with two binary operations, + and such that (R, +, 0) is an abelian
group; (R, , 1) is a monoid; and
a (b + c) = a b + a c,

(b + c) a = b a + c a

for all a, b, c.
(When there is no ambiguity we often write ab for a b.)

(2.3.5) Example. Can a set with one element be a ring? Yes.

(2.3.6) Example. Fix an integer m. Integers modulo m, written Zm , form a ring the ring of
modular arithmetic.
(2.3.7) Example. Let R be a ring. Then the n n matrices with entries in R, denoted Mn (R),
form a ring via the usual matrix algebra rules.
(2.3.8) Exercise. For S a set, let P (S) denote the power set. Note that the binary operations of
intersection and union are closed on P (S). Does (P (S), , ) form a ring? In fact if S is finite,
then (P (S), , ) is a ring (with playing the role of ; and , defined by A B = A B \ A B,
the role of +). Prove this, in case S = Z2 .
(2.3.9) Some workers do not require a ring to have an identity of multiplication. Both Lang and
Jacobson consider rings with 1, however.

Special elements in a ring

(2.3.10) The identity of the operation + in a ring is usually written 0 and called zero, even if the
ring does not contain Z.
(2.3.11) A unit in a ring is an invertible element.
(2.3.12) A zero-divisor in a ring is a nonzero element a such that there is another nonzero element
b with ab = 0.

12

2.3.1

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

Ring homomorphisms

(2.3.13) A ring homomorphism is a map

f :RS
between two rings such that f (ab) = f (a)f (b), f (a + b) = f (a) + f (b) and f (1) = 1.
(2.3.14) Example. There is a map f : Z Q given by f (n) = n.

(2.3.15) Example. There is a map g : Z M2 (C) given by g(n) = n12 .

(2.3.16) Example. The set map g : Z M2 (C) given by g(n) = diagonal(n, 0) is not a ring
homomorphism.
(2.3.17) Example. For any ring R there is a trivial homomorphism 1R : R R given by 1R (r) = r
for all r R.

(2.3.18) If for a ring homomorphim f : R S there is also a ring homomorphim f : S R such

that f f = 1R and f f = 1S , then f is a ring isomorphism.
If there is an isomorphism between two rings (f : R S, say) then the rings are said to be
isomorphic. We write R
= S.
(2.3.19) A ring isomorphism from a ring R to itself is called an automorphism.
Note that the set of automorphisms of a ring R forms a group under composition of maps, denoted
Aut(R).
(2.3.20) Example. The map : C C given by a + ib 7 a ib is a ring automorphism.

(2.3.21) Remark. It is an exercise to construct some simple examples of isomorphisms and automorphisms. We will give some interesting examples later, when we have constructed a few more
rings.
(2.3.22) The kernel of a ring homomorphism f : R S is the set
ker f := {r R | f (r) = 0}

(2.3.23) Example. The map f : Z Zm given by n 7 n = [n] (the class of n) is a ring

homomorphism, and ker f = mZ.
(2.3.24) If a ring homomorphism f : R S is an inclusion of sets, and the operations are the
same, then R is a subring of S.
(2.3.25) Example. For d Z the subset of M2 (Z) given by


a db
d
M2 (Z) := {
| a, b Z}
b a
is a subring.
(2.3.26) A subset R is a subring of ring S iff (i) the operations close in R; (ii) 1 R; and (iii) R
is closed under additive inverse.
Exercise: check this.
(2.3.27) Let f : R S be an inclusion of rings and t S. We write R[t] for the smallest subring
of S containing R and t.

2.3. RINGS

13

Note that the intersection of any collection of subrings of S is again a subring of S (exercise!).
It follows that R[t] is also the intersection of all the subrings of S containing R and t.
For every ordered tuple u0 , u1 , ...., un of elements of R (any n) we may form an element of S
by
X
u i ti
s=
i

If S is commutative then the set of all such elements (for all n N0 ) is a subring of S. This subring
is R[t] (since it contains R and t and is contained in R[t]).3
(2.3.28) Note that
Pn we do not say that s R[t] uniquely defines an ordered tuple from R. Indeed
if t is such that i=0 ui ti = 0 for some tuple with un = 1 (consider S = C, R = Z and t2 2 = 0
say) then it clearly does not. On the other hand, in such a case every element of the subring is
expressible as a polynomial of degree at most n
1 (since tn is).
Note that this setup includes our old friend Z[ d]. We will return to this again shortly.

Ideals
(2.3.29) Remark. An interesting way of constructing new rings from old is suggested by our kernel
example. Suppose r R and consider the set of all finite sums of elements of R of form arb:
X
ai rbi | ai , bi R}
RrR := {
i

(This is just an extension of our notation mZ.)

Notice that RrR behaves a bit like a ring itself: For x, y RrR we have x + y, xy RrR. But
note that 1 R may not be in RrR. (In fact if it is, then RrR = R.)

While we are introducing notation, lets have the following. For r R define r as the additive
inverse, i.e. r R such that r + (r) = 0. We then may abbreviate s + (r) as s r. Note that
(s r) = r s. (Check: r s + (s r) = r + (s) + s + (r) = r + (r) + s + (s) = 0.)
(2.3.30) Picking R and r R we can define an equivalence relation on R by
[x] = {y R | x y RrR}
We claim that the set of classes (denoted R/RrR) inherits a ring structure from R.
(2.3.31) In other words, for every R and r R we have another new ring R/RrR. (Although this
might often be the zero ring of course.) We automatically have a ring homomorphism
f : R R/RrR
given by f (x) = [x]. The kernel of this f is RrR.
(2.3.32) There is a very useful abstraction of this idea RrR as follows.
(2.3.33) An ideal in a ring R is a subset I such that (I, +) is a subgroup; and rx, xr I for all
x I and r R.
3 See

e.g. I.T.Adamson, Rings, modules and algebras, Oliver and Boyd.

14

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.3.34) Example. 3Z is an ideal in Z.

(2.3.35) Example. RrR is an ideal in R.
(2.3.36) Let I be an ideal of ring R. Then we define [r] R as the set of elements of the form
r +x, where x I. We define a relation on R by r I s if r s I. Note that this is an equivalence
relation. Thus each [r] is an equivalence class. (It is sometimes also useful to write r + I for [r].)
We write R/I for the set of equivalence classes.
Note that if rs = t then
(r + x)(s + y) = rs + ry + xs + xy = t + L
where L = ry + xs + xy, so that if x, y I then L I. It follows that we have a well-defined
multiplication on R/I given by [r][s] = [rs].
Also (r + x) + (s + y) = (r + s) + (x + y) so [r] + [s] = [r + s] is a well-defined addition. Thus R/I
is a ring.
(2.3.37) Theorem. [First ring isomorphism theorem] Let R be a ring. If : R S is a ring
homomorphism, then ker() is an ideal of R, (R) is a subring of S, and R/ ker()
= (R).
(2.3.38) Example. Consider the set M (Z) of 2 2 matrices of the form a1 12 + a2 where ai Z
and




1 0
0 1
12 =
, =
0 1
1 0
Does this form a subring of M2 (Z)?
Is f : M (Z) Z given by f : a1 12 + a2 7 a1 + a2 a ring homomorphism? If so, what is the
kernel? (...And in the form RrR?)
Is f2 given by f2 (a1 12 + a2 ) = a1 a2 a ring homomorphism?
Can we learn anything about M (Z) from this exercise?!
(2.3.39) We will study ideals further in Section 2.10.

Special kinds of ring

(2.3.40) A ring is an integral domain if it is commutative, has more than one element, and no
zero-divisors.
(2.3.41) Example. Z4 is not an integral domain, since 2 2 = 0; but Z is an integral domain.
(2.3.42) A field is a commutative ring that has more than one element, and such that every nonzero
element is invertible.
(2.3.43) Theorem. A finite integral domain is a field.
Proof. Let R = {r1 , r2 , ..., rl } be a finite integral domain. For any ri 6= 0 and j 6= k we have
ri (rj rk ) 6= 0, since rj 6= rk and there are no zero-divisors. Thus {ri r1 , ri r2 , ..., ri rl } = R and in
particular one of these is ri rj = 1. 2
(2.3.44) Theorem. Zn is a field iff n is prime.
Proof. (Only if): Suppose n = p1 p2 . Then p1 , p2 pass to zero-divisors in Zn .
(If): See later.

2.3. RINGS

15

(2.3.45) A matrix ring over a field is not in general a field. For example, the matrix ring M2 (Z2 )
is not a field.
(2.3.46) Exercise. (i) The matrix ring M2 (Z2 ) is finite. What is its order? Give an algorithm
for writing out all the elements; then do it.
(ii) Find a subring of M2 (Z2 ) that is a field, and which properly contains the natural image of Z2 .
Hints: Consider the subset

 
 
 

0 0
1 0
1 1
0 1
F4 =
,
,
,
0 0
0 1
1 0
1 1
(2.3.47) Suppose n Z is not a square. Then

Z[ n] = {a + b n | a, b Z}
This is a subring of C; and an integral domain.
(2.3.48) An ordered field (resp. ring) F is a field (resp. commutative ring) together with a total
order (F, ) such that a b implies a + c b + c for all c and a, b 0 implies ab 0.
(We will not focus on ordered fields here, but it is worth having the definition handy. In fact
there is another definition in common use axiomatising a notion of positive elements but it
is straightforward to show that it is equivalent to ours.)
(2.3.49) Examples: Q, R.
Note that an ordered field cannot have finite characteristic. And that the square of every
element is non-negative. Thus neither Z/pZ nor C is ordered.
(In fact every ordered field contains Q up to isomorphism (exercise).)
(2.3.50) Recall that a total order is used in the fundamental theorem of arithmetic (it is also used
in Euclids algorithm see later). We will need some replacement for this order in attempting to
prove such theorems for other rings.

The norm function

(2.3.51) Any element of Z[ n] as in (2.3.47) above can be written in a unique way as = a + b n

with a, b Z.
We define the norm
N () = |a2 nb2 |
(2.3.52) Lemma. (i) N () = 0 iff = 0; (ii) N () = N ()N ().
2

Proof. (i) a2 nb2 = 0 implies a2 = nb2 and hence n = ab2 . But this requires that n is a square.
(ii): Exercise. (Possible hint: cf. (2.3.25).)
Hereafter, when we omit a proof without comment, it is to be treated as an exercise.

(2.3.53) Lemma. is a unit in Z[ n] iff N () = 1.

(2.3.54) Example. In Z[ 3] we have N (7 4 3) = 1 so 7 4 3 are units.

The remainder of this subsection is an aside about a slightly more general setting for the norm
function, which is interesting but which we shall not need here.

16

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.3.55) Recall that a subset B of a k-vector space V is a basis if every element of V can be
expressed as a unique k-linear combination of elements from B.
Ordering the basis B = {b1 , b2 , ...}, a linear transformation t on V can be P
encoded as a matrix
m(t): One creates a column vector (b1 , b2 , ...)T and m(t) is such that t(bi ) = j m(t)ij bj that is

m(t)11
b1
b2 m(t)21
m(t)
=
..
..
.
.

m(t)12
m(t)22

t(b1 )
b1
...

...
b2 = t(b2 )
..
..
.
.

(2.3.56) Since a ring starts life as an abelian group, as does a k-vector space, there is a possibility
of endowing the ring with a scalar multiplication by k and hence endowing it with the property
of k-vector space. Indeed if our ring is a field it is already a vector space it is a vector space
over itself. It would also be a vector space over any subfield (we just have to restrict the ring
multiplication to the subfield).
If we have a vector spacethat is also a commutative ring along these lines (and specifically in
the sense of the example Q[ n] implicit in (2.3.51)) then ring multiplication by a given element
also looks like a linear transformation (on another element, and hence on the ring regarded as
a vector space). The matrix encoding this transformation depends on the basis. However the
determinant of the matrix does not.

2.4

Factorisation in integral domains

(2.4.1) Suppose R is an integral domain. Once again we define a|b to mean that ac = b has a
solution c R.
(2.4.2) Example. In Z[i] we have (1 + i)(2 i) = 3 + i, so 1 + i|3 + i.

(2.4.3) a, b R are associates if a = ub for u some unit. Association is an equivalence relation.

(2.4.4) p R is irreducible in R iff (i) p is not zero and not a unit; (ii) p = ab implies a or b a
unit.
(2.4.5) p R is prime in R iff (i) p is not zero and not a unit; (ii) p|ab implies p|a or p|b.

(2.4.6) Example. In Z an element is irreducible iff it is prime iff it is a prime number up to sign.
(2.4.7) Remark. In light of the case R = Z, we might like to suppose that the notions of primeness
and irreducibility are always interchangeable in an integral domain. In the next few paragraphs,
though, we shall see that this is asking to much, in general. Although they are indeed related...
(2.4.8) Lemma. In an integral domain R, prime implies irreducible.
Proof. Suppose for a contradiction that some prime p R obeys p = ab with a, b nonunits in R.
Then by primeness p|a or p|b say p|a WLOG. Then a = pc for some c R, so a = abc, thus
a(1 bc) = 0. Since a 6= 0 we have bc = 1, which contradicts that b is a nonunit. 2

(2.4.9) Ifa|b in R = Z[ n] then N (a)|N (b) as integers.

If a Z[ n] and N (a) is a prime number then a must be irreducible.

2.4. FACTORISATION IN INTEGRAL DOMAINS

17

(2.4.10) Example. In Z[ 5] the element = 1 + 5 is irreducible but not prime.

To see that is irreducible, note that N () = 6, thus any nonunit factorisation = would
have to obey N ()N () = 6. Thus the two norms are 2,3. But it will be clear that a2 + 5b2 = 2
has no solution for a, b Z.
Exercise: show is not prime.

(2.4.11) Theorem. In Z[ 2] every irreducible element is prime.

Before we prove this, we will need to parallel some more of the machinery we have for Z in this
more general setting.
(2.4.12) A GCD of a, b R is a common divisor c R such that if d is another common divisor
then d|c.
(2.4.13) Remark. It is not clear that GCDs exist in R in general, and indeed they may not.
But if a pair a, b have two GCDs then note that they are associates.

(2.4.14) Theorem. Suppose R = Z[ 2]. Then any pair a, b R has a GCD. If is this GCD
then it can be written in the form = ax + by for some x, y R.
Proof. This is an analogue of the situation for Z. There is an algorithm for computing the unknowns
starting from a, b in either case, called Euclids algorithm. In the Z-case this starts from the familiar
observation that a = qb + r has a solution with 0 r < |b|, where q is called the quotient, and r
the remainder. If course this just says that for every b there is an element r of the class of a in Z|b|
in the interval [0, |b|).

Here, with n = 2, we express b = l+m n and define b = lm n and M = bb = l2 (n)m2 .

Thus
a
ab
ab
=
=
b
bb
M

Write ab = t + s n, so that
t
s
a
n
=
+
b
M
M
Now let X, Y be the integers closest to the two ratios on the right, that is | Mt X| 1/2 and
similarly for Y . Next set

q =X +Y n
and r = a qb. We have

rb = ab qbb = ab qM = (t + s n) (X + Y n)M = (t M X) + (s M Y ) n

so
N (rb ) = (t M X)2 (n)(s M Y )2
so
N (r) =

N (rb )
N (rb )
t
s
=
= M ((
X)2 (n)(
Y )2 )
N (b )
M
M
M
M ((1/2)2 n(1/2)2 )

3
3
M = N (b) < N (b)
4
4

(Note well how this parallels the Z-case.)

(Note that this argument works also for n = 1, 2, 3, but not for n = 3.)

18

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

Using this idea we run an algorithm, first defining q1 , r1 by:
a = q1 b + r1

then q2 , r2 by
b = q2 r1 + r2
then qi , ri by
r1 = q3 r2 + r3
and so on. Note that N (r1 ) > N (r2 ) > .... Thus eventually some N (rk ) = 0. Let be the last
nonzero remainder = rk1 . Evidently |rk2 , so |rk3 and so on, so |b, so |a. Now if is
another common divisor then it divides all the remainders, so it divides . 2

(2.4.15) Theorem. In Z[ 2], if d|ab and the GCD of d and a is a unit, then d|b.
Proof. By (2.4.14) we can write a unit u = dx+ay. Thus b = buu1 = dbxu1 +abyu1. Evidently
d divides the RHS, so it divides b. 2

(2.4.16) Proof of (2.4.11): Say d is irreducible and d|ab in Z[ 2]. (We require to show, then, that
d|a or d|b.) Now let x be a GCD of d, a (recall that such a thing exists, by the Euclid algorithm
argument). Indeed set d = xe. By irreducibility of d, either x or e is a unit. If x is a unit then d|b
by (2.4.15). If e is a unit then d|x, and since x|a we have d|a. 2

UFDs
(2.4.17) A ring R is a unique factorisation domain (UFD) if it is an integral domain and
(i) every nonzero nonunit can be written as a product of irreducibles
x = p1 p2 ...pk
and (ii) if x = q1 q2 ...ql is another such factorisation then k = l and there is an ordering so that
pi , qi are associates for all i.
(2.4.18) Example. Z is a UFD: we can write 45 = (3).3.5 = 3.3.(5) = (5).(3).(3), and
1 is a unit.

(2.4.19) Theorem. Z[ 2] is a UFD.

Proof. The idea of this proof is to use the norm function to upgrade the proof of the Fundamental
Theorem of Arithmetic for use in this case.
To show factorisation, suppose for a contradiction that there is a nonzero, nonunit element
that cannot be written as a product of irreducibles. Then there is, in particular, such an element
with smallest norm N () (among all such elements). Obviously is not irreducible, so =
with , also nonunits. Thus N () = N ()N () with N (), N () > 1. Thus 1 < N () < N ()
and 1 < N () < N (). But then by assumption we can write , as products of irreducibles.
Thus we can do this for too a contradiction.
The proof of uniqueness is analogous to the case for Z, using Theorem (2.4.11) to allow us to
treat irreducibles as primes. 2

(2.4.20) Remark. The same

argumentshows that Z[i], Z[ 2], and Z[ 3] are UFDs. Z[ 5] is
not a UFD since 2.3 = (1 + 5).(1 5) are alternative factorisations of 6.

2.5. POLYNOMIALS

19

(2.4.21) Example. Let us write 30 Z[i] as a product of irreducibles. We have 30 = 2.3.5 for
starters, but also 2 = (1 + i)(1 i), so we can go further...
How can we be systematic about this?
If an integer can be written as a norm in Z (i.e. in the form n = a2 + b2 ), then n = (a + ib)(a ib).
The factorisation of 2 above comes from 2 = 11 + 12 ; and we also have 5 = 22 + 12 = (2 + i)(2 i).
Thus
30 = (1 + i)(1 i).3.(2 + i)(2 i)
Can we go further? (No.)

(2.4.22) Exercise. What happens if we form the quotient Z[ n]/rZ[ n] for r a unit; r an
irreducible; or r a prime?

2.5

Polynomials

See for example Kelarev4 [?, 3.4], or AndersonFuller5 [?, 1 Exercise 16], for a more general
setting for the following.
(2.5.1) Let R be a ring. A polynomial in indeterminate x, with coefficients in R, is a formal
expression of the form
f (x) = a0 + a1 x + a2 x2 + . . . + an xn =

n
X

ai xi

i=0

(any n N) with ai R. We say f (x) = g(x) if the coefficients ai agree for all i. We write R[x]
for the set of all such polynomials.

(2.5.2) Remark. Note that this is somewhat like Z[ d], except that we do not know d, so we
cannot eliminate x2 from expressions.
(2.5.3) We assume that you are familiar with real polynomial arithmetic. We define sum and product on polynomials in the usual way (treating the symbol + above as the usual + for polynomials),
making use of the operations in R in place of real arithmetic. For example
(a0 + a1 x) (b0 + b1 x) = a0 b0 + (a0 b1 + a1 b0 )x + a1 b1 x2
One sees that (R[x], +, 0) is an abelian group; and that (R[x], , 1) is a monoid. The operations
distribute appropriately, and so R[x] is a ring.
The usual polynomial arithmetic can be considered to be inherited from arithmetic in R,
treating x as an unknown, but an unknown from R. However one also usually assumes here that
R is commutative. If R is not commutative then the behaviour above is not right for x unknown
in R. However it is OK for x unknown central in R. (In practice we will take R commutative, so
there is no issue here.)
(2.5.4) Define R[x, y] = (R[x])[y] and so on.
4 Ring

constructions and applications. By Andrei V. Kelarev.

and categories of modules. By F W Anderson and K R Fuller.

5 Rings

20

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.5.5) Lemma. If R is a field then R[x] is an integral domain. The units are the nonzero constant
polynomials. Every nonzero polynomial is associated to a unique monic polynomial (a polynomial
with leading coefficient an = 1).
Proof. Note that R[x] is a commutative ring, and that it is not the
Pzero ring, and that the
Pconstant
i
polynomial 1 is the identity of multiplication. Suppose f (x) = ni=1 ai xi and g(x) = m
i=1 bi x ,
of maximal degree. We have
f g = an bm xn+m + ...
Since R is an integral domain an bm 6= 0. Thus f g 6= 0.
We leave the checking of units and associates as an exercise. 2
(2.5.6) Theorem. Let R be a field and f, g R[x] . Then there are q, r R[x] such that f = qg+r
with either r = 0 or degree r less than degree g.
Proof. Use polynomial division.
(2.5.7) Theorem. If R is a field then any pair f, g R[x] have a GCD h R[x]; and we can write
h = f u + gv for some u, v R[x].
Proof. A direct analogy of the Euclid algorithm from before. Note that it terminates this time
because the degree is decreasing. 2

The following are proved essentially as for the Z[ 2] case.

(2.5.8) Theorem. Suppose R is a field and consider R[x]. If the GCD of f (x), g(x) is a constant
polynomial, and f (x)|g(x)h(x) then f (x)|h(x).
(2.5.9) Theorem. If R is a field then every irreducible polynomial in R[x] is prime.
(2.5.10) Theorem. If R is a field then R[x] is a UFD.

2.6

Polynomials over Z

Now what about R[x] when R is not a field?

(2.6.1) Lemma. Z[x] is an integral domain. The units are 1.
Proof. Z is not a field, but in fact the same argument works here as in the case R[x] for R a field.
P
i
(2.6.2) If f =
i ai x Z[x] the content of f (denoted cf ) is the GCD of the coefficients.
Polynomial f is primitive if its content is 1.
(2.6.3) Example. Let f = 4x 6 and g = 10x2 + 15. We have cf = 2 and cg = 5.

(2.6.4) Remark. If f Z[x] \ Z is irreducible then it is primitive, since any non-constant f can
be expressed in the form cf f where f is non-constant primitive.
(2.6.5) Theorem. (Gausss Lemma) If f, g Z[x] are primitive then so is f g.

Proof. Hint: One can set this up as a proof by contradiction.

(2.6.6) Corollary: For any f, g Z[x] the content of f g is cf cg .

(2.6.7) Example. From our example (2.6.3) above we have cf g = 10. (Check this by brute force!)

2.7. IRREDUCIBLE POLYNOMIALS

21

(2.6.8) Remark. Note that each prime number p is an irreducible element in Z[x]; but no such
p is irreducible in Q[x]. OTOH, an element of the form x + a (a Z) is also irreducible in Z[x].
(How about in Q[x]?) What other irreducibles are there in Z[x]? Is x2 + 1 irreducible? Is x2 1?
Is x2 2?
(2.6.9) Theorem. If f Z[x] \ Z is irreducible in Z[x] then it is also irreducible in Q[x].

Proof. Hint: Aim for a contradiction.

(2.6.10) Theorem. (A) The irreducible elements of Z[x] are the numbers p with p a prime
number; and the primitive polynomials that are irreducible in Q[x].
(B) Every irreducible element in Z[x] is prime.
(C) Z[x] is a UFD.
Proof. Non-examinable.

2.7

Irreducible polynomials

(2.7.1) Theorem. (Factor theorem) Suppose R a field. If f R[x] and a R then (x a) is a

factor of f iff f (a) = 0.
Proof. The (only if) part is trivial. The (if) part uses polynomial division. 2
(2.7.2) Corollary: If R is a field then f R[x] of degree n has at most n roots in R.

(2.7.3) Theorem. (Fundamental theorem of algebra) Every f C[x] of degree > 0 has a root in
C.
Proof. Interestingly, this one is easier to prove using Analysis. 2
(2.7.4) Corollary: The irreducible polynomials in C[x] are the linear polynomials ax + b (note that
ax + b and x + ab are associates).
The irreducible polynomials in R[x] are the linear polynomials ax+b and the quadratics ax2 +bx+c
without real roots.
(2.7.5) Lemma. If R is a field and f R[x] has degree 2 or 3 and has no root in R, then it is
irreducible.
Proof. If f = gh one of g, h must have degree 1, giving a root in R. 2
(2.7.6) Example. Let f Z3 [x] be f = x3 + x2 + 2 . For x = 0 , 1 , 2 we have f (x) = 2 , 1 , 2
respectively. Thus f is irreducible.
Pn
(2.7.7) Theorem. (The rational root test) If f = i=0 ai xi Z[x] and r/s is a rational root
(with r, s coprime integers), then r|a0 and s|an .
Proof. Hint: Consider sn f (r/s) = 0.
(2.7.8) Example. Any rational root of f = x3 + x + 9 must be of form r/s with r|9 and s|1. That
is, r = 1, 3, 9 and s = 1. Trying all six cases one quickly finds that none gives a root. Since
the degree is 3, we deduce that f is irreducible in Q[x].

22

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.7.9) Theorem. (Eisensteins criterion) Let f =

number p such that
(i) p 6 |an
(ii) p|ai (i 6= n)
(iii) p2 6 |a0 .
Then f is irreducible in Q[x].

Pn

i=0

ai xi Z[x]. Assume that there is a prime

Proof. By (2.6.9) it is P
enough to show f irreducible in P
Z[x]. Suppose (aiming for a contradiction)
that f = gh with g = i bi xi of degree s > 0 and h = i ci xi with degree t > 0. Thus n = s + t,
P
and ai = ij=0 bj cij . Now p|a0 so p|(b0 c0 ), so p|b0 or p|c0 (but not both, since p2 6 |a0 ) say
p|b0 .
Now p|a1 so p|(b0 c1 + b1 c0 ) so p|b1 c0 ; and p 6 |c0 so p|b1 . Similarly p|bi for all i. But then p|an ,
giving a contradiction. 2
(2.7.10) Example. Taking p = 3 we see that 7x4 36x3 6x2 + 18x 12 is irreducible.

(2.7.11) Exercise. Is f = x10 + x9 + x8 + ... + x + 1 irreducible?

Obviously we cannot apply Eisenstein directly here (why not?). But note that f = f (x) is irreducible if f (x + 1) is irreducible... (Now consider p = 11.)

2.8

Fields of fractions

A commutative ring that is not a field fails to be a field by lacking inverses. Recently we have been
thinking about extending rings and fields by adding elements to them from rings which contain
them. Could we extend a ring to a field by adding inverses ? What if we do not have to hand a
larger ring from which to get these inverses?...
(2.8.1) Let R be an integral domain and consider the set R (R \ {0}). Define an additive binary
operation on this set by (r, s) + (t, u) = (ru + ts, su). This is closed and commutative. Is it
associative? Is there an identity element? An identity element would be an element (en , ed ) such
that (en , ed ) + (t, u) = (en u + ted , ed u) = (t, u) for all t, u. Such an element is (0, 1).
Define a multiplicative binary operation on this set by (r, s)(t, u) = (rt, su). This is closed and
commutative. Is it associative? Is there an identity element? An identity element would be an
element (in , id ) such that (in , id )(t, u) = (in t, id u) = (t, u) for all t, u. Such an element is (1, 1).
(2.8.2) The field of fractions of an integral domain R is the quotient of R (R \ {0}) by (r, s)
(t, u) if ru = st. Addition is represented by (r, s) + (t, u) = (ru + ts, su) and multiplication by
(r, s)(t, u) = (rt, su).
One needs to check that these are well-defined on classes.
Example: The field of fractions of Z is Q.
(2.8.3) Exercise. We have seen that R[x] is an integral domain. What is the field of fractions of
R[x]?
(2.8.4) Exercise. A field is an integral domain. What is the field of fractions of a field?

2.9. EXTENSION FIELDS

2.9

23

Extension Fields

(2.9.1) To say a field F extends a field R is just to say that R is a subfield of F (i.e. it is a subring
that is a field).
(2.9.2) Example. R is an extension field of Q.
(2.9.3) Let F be an extension field of a field R, and a1 , a2 , ..., an F . Define R(a1 , a2 , ..., an ) to be
the set of all elements of F obtained from elements of R and a1 , a2 , ..., an by repeatedly applying
the field operations: addition, multiplication, negation and multiplicative inverse.
This is a subfield of F (possibly all of it). It is called the subfield of F generated by R and
a1 , a2 , ..., an .
(2.9.4) We assume familiarity with real and complex vector spaces, bases and so on. (But please
do ask for a refresher if you like.) We note that these ideas all extend to vector spaces over a field.
(2.9.5) Exercise. Define the term basis of a vector space. How would this work over an arbitrary
field (e.g. Z2 )? What would go wrong if we tried to make it work over an arbitrary commutative
ring?
(2.9.6) The degree of a field extension F R, denoted [F : R], is the dimension of F as a vector
space over R.
(2.9.7) Example. [C :
R] = 2 since {1, i} is abasis for C regarded as a real vector space. For d
non-square we have [Q( d) : Q] = 2 since {1, d} is a basis.
It can be shown that [R : Q] = , that is, R can be regarded as a vector space over Q, but there
is no finite basis.
(2.9.8) Theorem. (Tower Law) Let L F R be field extensions. Then
[L : R] = [L : F ].[F : R]

Proof. The idea is to show that if {v1 , ..., vn } is a basis for F over R and {w1 , ..., wm } is a basis
for L over F then {vi wj }i,j is a basis for L over R. 2

(2.9.9)Example. Consider Q Q( 2) Q( 2, i). One can check that {1, 2, i, i 2} is a basis

for Q( 2, i) over Q.

2.9.1

Extending a field by algebraic elements

(2.9.10) Let F R be a field extension. Element a F is algebraic over R if there is a non-zero

polynomial f R[x] with f (a) = 0 in F .
Otherwise a is transcendental over R.

(2.9.11)
Example. (I) 2 R is algebraic over Q, since it is a root of f = x2 2 Q[x].

(II) 5 2 is algebraic over Q, since it is a root of f = x5 2 Q[x].

(III) i Cpis algebraic over Q, since it is a root of f = x2 + 1 Q[x].

(IV) a =
7 1 is algebraic over Q, since it is a root of f = x4 + 2x2 6 Q[x].
(V) R is transcendental over Q (Lindemann, 1882). How might you prove this?

24

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.9.12) Exercise. What can we say about algebraic extensions of the field Z2 ?
(2.9.13) If a is algebraic over R then the minimal polynomial of a over R is the unique monic
polynomial f R[x] of least degree with f (a) = 0.

(2.9.14) Example. Element 2 is a root of x4 4 and 2x3 4x, but its minimal polynomial is
x2 2.

(2.9.15) Remark. Note that the minimal polynomial exists since by the algebraic assumption
there is a polynomial g R[x] with g(a) = 0. But if the leading coefficient is gn then (1/gn )g is
monic with the same root. Among such polynomials will be at least one with least degree. If it is
not unique then there are two of this degree, but then g g also has root a and lower degree
leading to a contradiction.
(2.9.16) Lemma. Suppose F R is a field extension, and a F algebraic over R with minimal
polynomial f . Then (i) f is irreducible in R[x]; (ii) If g R[x] and g(a) = 0 then f |g.
Proof. (i) If f factorises as gg then one of g(a), g (a) must be zero. Rescaling leads to a polynomial
of smaller degree a contradiction.
(ii) Division gives g = qf + r. Then r(a) = 0, so in fact r(x) = 0 (else it contradicts the minimality
of f . 2
(2.9.17) Theorem. Suppose F R is a field extension, and a F algebraic over R with minimal
polynomial f . Then f is the unique monic irreducible polynomial in R[x] with f (a) = 0. (I.e.
irrespective of degree.)
Proof. Suppose g has these properties. Then f |g, so g = f h for some h. But then h is constant
by irreducibility, and this constant is 1 since f and g are both monic. 2

(2.9.18) Example. n 2 is algebraic over Q since a root of xn 2. This polynomial is monic

irreducible in Q[x] (by Eisenstein with p = 2, for example), so it is the corresponding minimal
polynomial.
(2.9.19) Proposition. Suppose F R is a finite field extension. Then every element of F is
algebraic over R.
Proof. Let a F . Then all powers of a are in F . But these cant all be linearly independent over
R. 2
(2.9.20) Theorem. Suppose F R is a field extension, and a F algebraic over R with minimal
polynomial fa of degree n. Then R(a) has basis {1, a, ..., an1 } as a vector space over R.
In particular [R(a) : R] = n, so R(a) is a finite extension field of R.

(2.9.21) Example. a = 3 2 has fa = x3 2 over Q. Thus Q(a) has basis 1, a, a2 over Q. Thus
Q(a) = { + a + a2 : , , Q} and [Q(a) : Q] = 3.
Proof. Let fa = xn + an1 xn1 + ... + a0 x0 . The powers 1, a, a2 , ..., an1 are linearly independent
over R by the minimal condition. Let E be the R-subspace of F spanned by these. We will next
show by induction that an+k E for all k 0.
We have fa (a) = an + an1 an1 + ... + a0 a0 = 0, so the base case k = 0 is true. Now suppose
true up to level k 1 (the inductive assumption). Multiplying by ak we have
an+k = an1 an+k1 an2 an+k2 ... a0 ak

2.9. EXTENSION FIELDS

25

But everything on the right is in E by assumption, and hence so is an+k .

Thus E contains all positive integral powers of a.
It follows from this that E is closed under products, so it is a subring of F . Next we show
that it is a subfield. Suppose b E . Then let b = b0 a0 + ... + bn1 an1 be an expansion in the
spanning set. Thus b = g(a) where g = b0 x0 + ... + bn1 xn1 . Since g(a) 6= 0 we know that fa 6 |g.
As fa is irreducible, it and g have a GCD which is a unit. Thus there are polynomials u, v such
that fa u + gv = 1. Thus g(a)v(a) = 1. Thus b1 = v(a) E.
Finally we show that E = R(a). The inclusion E R(a) follows from the definition, and the
reverse inclusion holds since E is a field which contains R and a. 2

Throughout this section we have had the assumption that F R is a field extension, and
we have considered the nature of elements of F in relation to R. This contrasts with our earlier
construction of fields of fractions, where we wanted to extend a (certain kind of) ring without
having a pre-existing structure to extend it into.
If our subfield is a subfield of the complex numbers then of course every polynomial over the
subfield has a root over C, so extending the subfield by a root of a polynomial is the same as
extending by a concrete complex number. In general though, we could start with a field and a
polynomial f over that field, and it is not a given that there is an extension field containing a
root of f .
Except that...
(2.9.22) Theorem. (Kroneckers Theorem) A polynomial f in a field F has a root in an extension
field.
Proof. We may assume that f = f (x) does not have a linear factor in F [x] since otherwise that
factor would give the root. Let g(x) denote a factor of f (x) which is irreducible. Let E =
F [x]/gF [x]. Since g is irreducible this ring is a field containing F , and hence an extension field of
F . The representative x F [x] of the element x E satisfies f (x) = h(x)g(x) 0 (mod g(x), as
it were), so f (x ) = h(x ) g(x ) = 0 . 2
(If the learning curve is a bit steep here, first recall that the basic model for something like
F [x]/gF [x] is Zp = Z/pZ (p prime); and then have a look at some of the exercises and hints in
section 2.11, such as Theorem (2.11.2).)
(2.9.23) Exercise. (Optional) Show that the smallest extension field of Z2 in which p1 (x) =
x4 + x + 1 Z2 [x] has a zero is isomorphic to the smallest extension field of Z2 in which p2 (x) =
x4 + x3 + 1 Z2 [x] has a zero.

2.9.2

Remarks on Kronecker

(2.9.24) Recall the fundamental theorem of algebra. From this we see that, while there are lots
of non-algebraically closed fields, the fields Q and so on contained in C can be extended to an
algebraically closed field.
We now see that the Kronecker Theorem says something similar (but more general). If we start
with an arbitrary field F and form polynomials, then either F is algebraically closed or eventually
we will find a polynomial without a root in F . But then we can form an extension field of F , via this

26

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

polynomial, which does contain a root of F . Iterating this process to closure (i.e. exhaustively)
we will extend F to an algebraically closed field.
On the other hand note (cf. e.g. Fulton [?]):
(2.9.25) Let F be any field. Let f1 , f2 , ..., fn be a set of irreducible monic polynomials in F [x].
Then f1 f2 , ...fn + 1 is not divisible by any of these. Thus there are infinitely many distinct
irreducible monic polynomials.
(2.9.26) It follows that the algebraic closure of any field F is infinite.

2.9.3

Geometric constructions

(2.9.27) We say a real number a is constructible if starting with two marked points in the plane at
distance 1 apart, we can construct two marked points with distance |a| between them, using only
the following operations:
(i) draw line through any pair of marked points;
(ii) add the distance between these to our set of constructed constructible distances;
(iii) draw a circle of constructed radius centred at a marked point (e.g. the circle defined by that
point and another marked point);
(iv) mark any point of intersection of lines and circles.

(2.9.28) Example. Any number in Q is constructible. 2 is constructible.

(2.9.29) Exercise. (Optional) What other kinds of real numbers are constructible?
(2.9.30) Lemma. The set Sc of constructible numbers is a subfield of R. If a Sc then so is

p
|a|.

Proof. Optional exercise. (The key point is that we have embedded the real line in the real plane.)
(2.9.31) Theorem. (Wantzel 1837) If a constructible, then it is algebraic over Q and the degree
of its minimal polynomial over Q is a power of 2.
Proof. Optional exercise.

2.10

Ideals

Let R be a ring. We can partially order the ideals of R by inclusion. Obviously R itself is the top
of this order, but if we restrict the order to the proper ideals then there could be many maximal
elements in the order.
For example, 36Z 18Z 6Z 3Z, but we cant go any further without hitting Z, so 3Z is a
maximal element among the proper ideals of Z. Indeed pZ is a maximal element iff p prime.
(2.10.1) In general we call the maximal elements of this order the maximal ideals of a ring.
(2.10.2) A proper ideal P in a commutative ring R is a prime ideal if whenever ab P then either
a P or b P .
(We will give a definition for prime ideal in an arbitrary ring in (2.10.21).)
(2.10.3) Example. The maximal ideals of Z are also prime. (We will see later that maximal ideal
implies prime ideal in any commutative ring R.)

2.10. IDEALS

27

Ideal arithmetic, towards Principal ideal domains

(2.10.4) Exercise. Show that the intersection of two subgroups of a group is a group.
Answer: Let H, H G be groups. If g, f H H then g, f H, H so gf H, H , so
gf H H . Thus multiplication closes in H H . Evidently the identity element e of G lies in
H and H and hence in H H . Finally if g H, H then g 1 H, H so H H also has inverses.
2
(2.10.5) Let R be an abelian group and S a subset. Let us write hSi = hSiR for the intersection
of all subgroups of R containing S. This is called the abelian group generated by S in R.
(2.10.6) The group hSi is the set of finite sums of elements of S and their additive inverses.
Example: h3iZ = 3Z = h3, 6iZ .

(2.10.7) Recall that an ideal in a ring R is a subgroup J such that rar J for all r, r R and
a J.
(2.10.8) Note that if J is an ideal of R then J = hJiR .

(2.10.9) Lemma. The intersection of two R-ideals is an R-ideal.

Proof. If A, B ideals and x A, B then rx A, B (and so on). Closure under addition is also
clear. 2
(2.10.10) Note that if {Ji }iT is a set of R-ideals then i Ji is an R-ideal.

J_2

J_1

J_i

J_3

(2.10.11) Example. Let S R a subset, and let {Ji }iT (S) be the subset of ideals of R such that
Ji S. Then
(S) := i Ji
is an ideal.
(2.10.12) This (S) is the smallest R-ideal containing S. It is called the ideal generated by S.
(2.10.13) Let S = {s1 , s2 , ..., sn } R. Elements of (S) = (s1 , s2 , ..., sn ) take the following form.
Firstly we have all constructs of form rsi r . Then we have all finite sums of such constructs:
XX

R; si S}
| riji , rij
riji si rij
(s1 , s2 , ..., sn ) = {
i
i
i

That is, (S) = h{rar | r, r R; a S}iR .

ji

28

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(2.10.14) Let R be a ring. For A, B subsets of R we define another subset of R by

AB := {a1 b1 + a2 b2 + ... + an bn | n N, ai A, bi B}
(where the empty sum is zero). Note that this is the finite additive closure of the set of elements
of the form ab with a A and b B.
(2.10.15) Note that if A, B are ideals then so is AB. (See e.g. Jacobson I 2.5.)
(2.10.16) Suppose A, B are R-ideals. What is (AB)A?
(2.10.17) Exercise. Whats is A(BA)? How are they related?
(Answer: (AB)C = A(BC) see e.g. Jacobson I 2.5.)
(2.10.18) Let R be a ring and S a subset of R. The subset of R generated by S is the set RSR.
(2.10.19) Let R be a ring. For A, B subsets of R we define another subset of R by
A + B := {a + b | a A, b B}
If A, B are R-ideals then so is A + B. (See also, e.g., ZariskiSamuel [?, III.7].)

(2.10.20) Exercise. What are A B, A + B and AB in each of the following cases?:

(i) A = 2Z and B = 3Z;
(ii) A = 2Z[x] and B = xZ[x];
(iii) A = 2Z[x] and B = 2xZ[x].
(2.10.21) A proper ideal P in an arbitrary ring R is a prime ideal if, for ideals A, B, whenever
AB P then either A P or B P .
(Note that this definition contains the commutative case.)

Principal ideal domains

(2.10.22) An ideal in a commutative ring R is principal if it can be expressed in the form dR for
some d R.
(2.10.23) We have repeatedly used the construction dR for ideals in commutative rings. For
example 3Z is an ideal in Z.
(2.10.24) Lemma. Every ideal in Z is principal.
Proof. Let I be an ideal in Z and d the smallest positive element. Suppose for a contradiction that
I \ dZ 6= , and let d be the smallest positive element. Then d ld [1, d 1] for some l. But
d ld I, contradicting that d is smallest. 2
(2.10.25) A principal ideal domain (PID) is an integral domain such that every ideal is principal.

(2.10.26) Exercise. Show that every ideal in Z[ 2] is principal.

(2.10.27) Exercise. Show that we can choose a d so that Z[ d] is not principal.

(2.10.28) Exercise. Show that the ideal of Z[x] generated by 2 and x (i.e. the smallest ideal
containing these two elements) is not principal.
(2.10.29) Lemma. If r is an irreducible element in a PID R then rR is a maximal ideal.

2.10. IDEALS

29

Proof. Else rR aR R for some a, so r = ab for some nonunit b, a contradiction. 2

(2.10.30) Lemma. A maximal ideal M in a PID R has the property that ab M implies a or
b M.
Proof. For, suppose ab M but a, b 6 M . Then M + aR = {m + ar | m M, r R} and M + bR
are both ideals properly containing M . Since M is maximal this says that M + aR = R = M + bR.
We have aRbR abR M , so R = R2 = (M + aR)(M + bR) M 2 + aRM + M bR + aRbR M
a contradiction. 2
(2.10.31) Lemma. A PID is a UFD.
Proof. First we show that any element x of a PID R can be factorised into a list of irreducible
factors. Clearly if x cannot be factorised then it is irreducible, so the only obstruction is if in fact
it can be factorised but this process does not terminate! We rule this out as follows:
Suppose R is an integral domain, and (x =)r1 , r2 , r3 , ... is a sequence of elements such that
ri |ri1 .
(The example one might have in mind here is something like 100, 50, 10, 5, 2, 1.)
It follows that ri R ri1 R. Let I be the union of all these ideals.
(This union might be over infinitely many ideals, but every element of I is an element of some ri R.
Just as Z is infinite, but every element is a finite number.)
If R is a PID then there is an r such that rR = I. Thus there is some i with r ri R. At this point
I = rR = ri R (we have just shown that rR ri R, and rR ri R by construction), and rj R = rR
for all j i (so that all subsequent rj s are associates).
In particular, it is not possible to have an infinite sequence of ri s as above that are all non-associate.
(In our example this just says that every such positive integer sequence must terminate.)
But an infinite irreducible factorisation x = p1 p2 ... would give rise to an infinite sequence r1 = x =
p1 p2 ..., r2 = p2 p3 ..., with ri |ri1 . Thus there can be no such factorisation.
Now we set out to show uniqueness. We will need to note a couple of facts.
(2.10.32) Lemma. In a PID, irreducible implies prime.
Proof. First note that r irreducible implies rR maximal, by Lemma 2.10.29. But any maximal
ideal M has the property that ab M implies a or b M , by Lemma 2.10.30.
Since rR is maximal it has this prime ideal property. 6 That is, whenever r|ab, then r|a or
r|b. Thus r is prime.
2
(2.10.33) (Now back to complete the proof of (2.10.31))
Suppose p1 p2 ...pr = q1 q2 ...qs are two irreducible (hence prime) factorisations of x R. By primeness p1 must divide some qi . Hence p1 = uqi for some unit u (by irreducibility). Reorder so that
q1 = qi . We then have uq1 p2 ...pr = q1 q2 ...qs so up2 ...pr = q2 ...qs . Up to the unit, this is analogous
to the initial equality except with one fewer irreducible on each side, so we can iterate, pairing
each pi with a qi up to a unit. This establishes uniqueness. 2
(2.10.34) Example. Since Z is a PID, it is a UFD. (But of course we already knew this.)
(2.10.35) Exercise. Can you think of a UFD that is not a PID?
(Hint: Think about Z[x].)
6 An ideal M is prime if whenever A and B are ideals and AB is a subideal of M then either A or B is a subideal
of M .

30

2.11

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

More Exercises

(2.11.1) For what nonsquare values of d is the set

Id = {a + b d | a, b Z; a b even}

an ideal in Z[ d]?

HINTS: Suppose
= a + b d and = e + f d both in Id . We have (a + b d) + (e + f d) =
(a + e) + (f + b) d and a + e f b = (a b) + (e f ), so we have closure under addition for
any d.

Now for Id and Z[ d] we have (a + b d)(e + f d) = ae + dbf + (af + be) d and

ae + dbf (af + be) = e(a b) + f (db a) The first part is even for any e, and the second part is
even for any f provided that d is odd.
(2.11.2) Theorem. Let F be a field. Then F [x]/gF [x] is a field if g is an irreducible polynomial
in F [x].
Prove it. (What about the only if version?)
HINTS: Let h F [x] be a representative of h F [x]/gF [x] . Since g is irreducible, the GCD
(g, h) = 1. Thus there is a u, v such that gu + hv = 1 in F [x]. But the image of this in the quotient
is h v = 1 , so there is an inverse v of h .
(2.11.3) CLAIM: Let F be a field and g a nonconstant polynomial in F [x]. The map F
F [x] F [x]/gF [x] given by f 7 f x0 + 0x1 + ... 7 f is an injection.
Prove it.
HINTS: Suppose g = 1 + x2 for example. Then 1 = {1, 1 + g, 1 + 2g, ..., 1 + (2 + x)g, ...} Note
that no other constant polynomial besides 1 occurs.

2.12

More revision exercises

(2.12.1) Find a GCD of a = 5 + 14i and b = 4 + 7i in Z[i].

HINTS: We could start by trying to find a quotient-remainder formula: a = bq +r (say). Noting
that a has the bigger norm, we compute a/b, which will eventually give us a/b = q + r/b:
5 + 14i 4 7i
78 91i
6 7i
a
=
=
=
b
4 + 7i 4 7i
65
5
5
This is a/b = (1 i) + (1 2i)/5 (by first rounding to the nearest integers, then computing the
correction) so
a = (1 i)b + ((1 2i)/5)b

The second summand (the remainder) is 51 (1 2i)(4 + 7i) = 15 (4 + 14 + 8i + 7i) = 2 + 3i (of

course we knew the denominator
would cancel, since a (1 i)b Z[i]).

The norms are N (2 + 3 1) = |22 (1)32 | = 13, and N (5 + 14i) = |52 (1)142| =big, and
N (4 + 7i) = |(4)2 (1)72 | =big-ish. This just checks that the remainder has small enough
norm (as the rounding method is designed to ensure have a think about how it does this).
We next compute
4 + 7i 2 3i
13 + 26i
b
=
=
= 1 + 2i
2 + 3i
2 + 3i 2 3i
13

2.13. HOMEWORK EXERCISES

31

Normally we would have to keep going, but since this quotient is in Z[i] there is no remainder, and
we see that 2 + 3i|b. Thus 2 + 3i also divides a and is a common divisor.
(2.12.2) Find as many subrings as possible of Q.
(2.12.3) If : R S is a ring homomorphism, what can we say about (0)?
(2.12.4)
Determine
are ring homomorphisms:
which of the following
mappings
(1) : Z[ 2] Z[ 2] defined by (a + b 2)
=
a

b
2for a, b Z.

by 2 (a + b 2) =b + a 2 for a,
(2) 2 : Z[ 2] Z[ 2] defined

b Z.
(3) 3 : Z[ 2] M7 (Z[ 2]) defined by 3 (a + b 2) = (a + b 2)17 for a, b Z (where 17 is the
identity matrix).
(2.12.5) (I) Let I be an ideal in a ring R. Show that the multiplication in the factor ring R/I is
well-defined.
(II) Give the multiplication table for the factor ring Z/3Z.
(2.12.6) For y R define Ay = {f Q[x] | f (y) = 0}. Under what circumstances is Ay an ideal
in Q[x]?
(2.12.7) Write x4 2 as a product of irreducible polynomials over each of Q[x], R[x], and C[x].
(2.12.8) Determine which of the following polynomials are irreducible over Q:
(a) 2x3 + 5x2 2x + 3.
(b) x4 + 55x2 + 1210x.

2.13

Homework exercises

Here we write det M for the determinant of a square matrix M .

(2.13.1) Show that a sufficient condition for a matrix M in the ring M2 (Z) to be a unit in M2 (Z)
is det M = 1.
(2.13.2) Consider the subset S x of matrices in M2 (Z) such that all four matrix entries are nonzero.
Give an example of an element of S x :
(i) that is a unit in M2 (Z);
(ii) that is a non-unit in M2 (Z) but a unit in M2 (Q); and
(iii) that is a non-unit in M2 (Z) and in M2 (Q).
(2.13.3) Show that for any unit U in the ring R = M2 (Z) the map
fU : R R
given by fU (X) = U XU 1 is a ring homomorphism from R to itself.
Construct an inverse map to fU .
(2.13.4) There are infinitely many elements X of M2 (Z) with det X = 0. Recall from lectures that
for each choice of d Z we have a subring


a db
M2d (Z) = {
| a, b Z}
b a

32

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

(i) Prove carefully that, for any choice of d, M2d (Z) is a ring.
(ii) For what choices of d does the ring M2d (Z) contain nonzero elements X with det X = 0?
(iii) For what choices of d is M2d (Z) a commutative ring?

(2.13.5) Consider
a fixed but arbitrary d Z, and consider the ring R = Z[ d]. Suppose x R
obeys x = a + db with a, b Z. For what values of d does x determine a unique values for the
pair (a, b) in
this way? In the other cases, what can be said about the set of pairings (a, b) such
that x = a + db, for any given x.
(2.13.6) For each d determine whether the map

fd : M2d (Z) Z[ d]
given by

a
b

db
a

7 a +

db is a ring homomorphism.

(2.13.7) Prove that M22 (Z) is a UFD.

Answers
1. The inverse of M if it exists is easily verified to be the transpose of the matrix of cofactors
divided by det M . The transpose of the matrix of cofactors is integral by construction. 1 divide
every integer.


1 2
2. Examples: (i)
1 3


1 2
(ii)
1 4 
1 2
(iii)
1 2
3. fU (a + b) = U (a + b)U 1 = U (a)U 1 + U (b)U 1 = fU (a) + fU (b) (via distributivity);
fU (a.b) = U (a.b)U 1 = U (a)U 1 .U (b)U 1 = fU (a).fU (b)
Inverse is fU 1 .
4. (i) Observe that it is a subset. Now check closure and inverses.
(ii) d a perfect square.
(iii) any.
5. (i) d not a perfect square;
(ii) There are infinitely many (with a, b related by a simple formula).
6. Yes. To see this check fd (ab) = fd (a)fd (b) and fd (a + b) = fd (a) + fd (b) for each d (routine
calculations, which you should do!, with d left as an arbitrary but fixed number).

Homework exercises 2
(2.13.8) For n > m consider the set map
: Mn (Z) Mm (Z)

2.13. HOMEWORK EXERCISES

a11
a21
..
.

am1

.
..
an1

a12
a22
am2
an2

...
...

a1m
a2m

... amm
...

anm

33
...
...

a1n
a2n
..
.

... amn
..
.
...

ann

a11

a21

7
..

am1

a12
a22

...
...

a1m
a2m
..
.

am2

... amm

For which values of n, m is this a ring homomorphism (if any)? Give reasons for your answer.
(2.13.9) Show that the intersection of two subgroups of a group is a group.
(2.13.10) For R a ring and A a subset of R, let s(A) denote the set of all subrings of R that
contain A (including R itself). Show that the intersection of all these subrings is itself a subring
of R.
This intersection subring is called the ring generated by A in R.
Suppose that 0 6= 1 in R. Show that the sets , {0} and {1} all generate the same ring in R.
Construct a ring R with 0 6= 1 such that the ring generated by in R is
(i) isomorphic to Z;
(ii) not isomorphic to Z.
(2.13.11) Consider the polynomial f = x2 + 1. Regarded as a polynomial over which of the
following coefficient rings is this polynomial irreducible? (i) Z; (ii) R; (iii) C; (iv) Z2 . Give reasons
for your answers.
(2.13.12) Explain why the polynomial x4 7 is irreducible over Q, quoting any theorems you use.
(2.13.13) Give the multiplication table for the ring Z/4Z.

Answers
8. Define A Mn (Z) (any n > 1) by a1n = an1 = 1 and all other entries zero. Then (A) = 0
(any m < n) but (A2 ) 6= 0. Thus never a homomorphism. (Other formulations are possible.)
9. Let G, G be the subgroups. If a, b G G then a, b G, G so ab in G, G (since they are
groups), so ab GG , so the operation is closed in GG . It is associative by restriction similarly.
The identity lies in both, hence in the intersection. Inverses lie in the intersection similarly.
10. Suppose elements a, b lie in the intersection. Then they lie in every subring. Now proceed
similarly to above.
The smallest subring containing is simply the smallest subring. But every ring contains 0, 1,
so this subring does. Thus it is the same as the subring generated by 0, or 1.
(i) R (other answers are possible);
(ii) Z2 .
11. This is a quadratic, so it is irreducible iff it had no root in the ring. Thus: (i) irreducible,
since every element of f (Z) is positive; (ii) irreducible, since every element of f (R) is positive; (iii)
not irreducible (i is a root); (iv) not irreducible (1 is a root).
12. Use Eisensteins criterion with p = 7. (Or could determine a complete factorisation over C,
and note that none of the roots lie in Q C.)
13. Elements of Z/4Z are {[0], [1], [2], [3]}. Multiplication is then just as in the usual mod.4
arithmetic.

34

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

Homework exercises 3
(2.13.14) Determine if the map X : Z[x] Z[x] given by f (x) 7 f (x+3) is a ring homomorphism.

(2.13.15) Explain what is meant by Q( 7).

(2.13.16) Suppose K F is a field extension Explain what it means for a K to be algebraic
over F .

(2.13.17) Prove that 7 is irrational.

p
(2.13.18) Determine a basis of Q(
5 11) over Q.
(2.13.19) Construct an irreducible quadratic polynomial in Z2 [x] (where Z2 = Z/2Z).

(2.13.20) Let f (x) = 2x3 + 3x2 + 9x + 12 Z[x]. Is this polynomial primitive? State Eisensteins
criterion, and use it to determine if f (x) is irreducible over Q.
Compute the polynomial x3 f (1/x). Can Eisensteins criterion be applied directly to this polynomial? Is it irreducible (give reasons for your answer)?
(2.13.21) Let g(x) = x3 + 3x + 9 Z[x]. This polynomial is irreducible over Q, but why can we
not use Eisensteins criterion to show this directly?
(2.13.22) Determine which of the following polynomials are irreducible over Q, explaining your
method in each case.
(i) f0 (x) = 3x4 + 15x2 + 10;
(ii) f1 (x) = 3x9 + 15x6 + 10;
(iii) f2 (x) = x2 + x + 2;
(iv) f3 (x) = f4 (x + 3), where f4 (x) is an irreducible polynomial.
(2.13.23) Let R be a ring and I, I ideals of R. Show that I I is an ideal of R.

Answers
14. We need to check that X(f g) = X(f )X(g), and X(f + g) = X(f ) + X(g). We have
X(f g) = X(f (x)g(x)) = X((f g)(x)) = (f g)(x + 3) = f (x + 3)g(x + 3) = X(f (x))X(g(x))
X(f +g) = X(f (x)+g(x)) = X((f +g)(x)) = (f +g)(x+3) = f (x+3)+g(x+3) = X(f (x))+X(g(x))
(Note also
that X(1) = 1 and X(0) = 0.)

15. Q( 7) is thefield of elements of form a + b 7 (a, b Q). This is the smallest field extension
of Q containing 7.
16. a K is algebraic over F if there
is a polynomial in the ring F [x] with root a.
17. Suppose for a contradiction that 7 = a/b with a, b Z coprime. Then 7b2 = a2 . But then
7|a2 and hence 7|a, since 7 is prime. But then 72 divides a2 and hence 7 divides b2 . So, similarly,
7 divides b, contradicting
coprimeness.
p

18. Write a =
5 11. Then a2 = 5 11 and (a2 + 11)2 = 5 so a is algebraic with minimal
polynomial fa = (a2 + 11)2 5 = a4 + 22a2 + 116 (note that this is irreducible: there is no
rational root, or indeed real root, since the coefficients are positive;pthus the only possibility is

a factorisation into two quadratics; but the roots over C are ai = 5 11 and one readily

2.14. MORE (OPTIONAL) RINGS AND EXERCISES

35

checks that no quadratic factor is integral).

A basis is therefore {1, a, a2 , a3 }.
19. Try f (x) = x2 + x + 1. Since it is quadratic, it is enough to show that it has no roots in Z2 .
We have f (0) = f (1) = 1, so done.
20. The polynomial
is primitive since the first two coefficients are coprime. Eisensteins criterion
Pn
says f = i=0 ci xi is irreducible over Q if there is a prime p such that p 6 |an and p|ai otherwise
and p2 6 |a0 .
Thus f is irreducible by Eisenstein with p = 3.
g = x3 f (1/x) = 12x3 + 9x2 + 3x + 2 and Eisenstein cannot show irreducibility directly since there
is no suitable prime.
g is irreducible since if it factorised this would induce a factorisation of f , contradicting its irreducibility.
21. There is no suitable prime.
22. (i) irreducible by Eisenstein with p = 5
(ii) ditto
(iii) irreducible by rational root test (Suppose r/s a root, with r, s coprime: then (r2 +rs+2s2)/s2 =
0, so r2 = s(r + 2s) so s = 1 (else contradicts coprimeness), giving possible roots 1, 2. These
all fail.)
(iv) irreducible since any factorisation induces a factorisation of f4 .
23. Need to show J = I I an abelian group closed under the R action. Suppose j, k J. Then
j, k I so j + k I, and similarly for I , so j + k J. For r R we have rj I, I (since these
are R-ideals) so rj J. Similarly for jr.

2.14

More (Optional) Rings and Exercises

(2.14.1) Let R1 , R2 be rings. Define operations on R1 R2 by

(r1 , r2 ).(s1 , s2 ) = (r1 s1 , r2 s2 )
(r1 , r2 ) + (s1 , s2 ) = (r1 + s1 , r2 + s2 )
Show that R1 R2 can be made into a ring using these operations.
(The new ring is called the direct product or direct sum of R1 , R2 .)
(2.14.2) To form a product of three rings, Ra , R1 , R say, we need to address some practicalities
of notation. For one thing the indexing set I = {a, 1, } is not necessarily ordered (that is, we
may not be given an order on the set, and there may not be a natural way to assign one). For
another, the Cartesian product is not strictly associative. Thus the underlying set of the product
is not strictly Ra R1 R (indeed it is not necessarily clear what this means). A convenient
resolution is to consider R = iI Ri to be the set of indexed triples of elements from the three
rings. That is, if r R then it contains, for each i I, a component ri ; and the collection of these
components determines r.
(2.14.3) Show that the above direct product construction extends in the obvious way to any finite
set of rings {Ri | i I} (with I a finite indexing set).
(2.14.4)
Q The construction may be extended still further to infinite indexing sets I (we write
R = iI Ri ), but where R is such that r R implies that only finitely many of the ri are

36

CHAPTER 2. RINGS, POLYNOMIALS AND FIELDS

different from the zero element of the ring Ri . Explain why the given pointwise addition and
multiplication close on R constructed in this way.

2.14.1

Group rings and skew group rings

(2.14.5) If R is a ring and S a semigroup (resp. monoid, group) then the semigroup (resp.Pmonoid,
group) ring R[S] is defined as follows. The underlying set is the set of all formal sums sS rs s
where rs R and only finitely many of these are nonzero. (See e.g. Kelarev [?, 3.2].) Addition is
X
X
X
(
rs s) + (
rs s) =
(rs + rs )s
sS

and

sS

X
X
(
rs s) . (
rs s ) =
s S

sS

sS

(rs rs )ss

s,s S

(2.14.6) Suppose R is a ring and G a group of automorphisms on R; and write rg for the image
of r R under automorphism g G. The skew group ring R G is the same set and additive
structure as R[G]; this time made into a ring by
(rg).(r g ) = ((r(r )g

)(gg ))

(2.14.7) The fixed subring of G on R is

RG := {r R | rg = r g G}

2.14.2

Monoid-graded algebras

(See e.g. Karpilovsky [?, 22].)

(2.14.8) An algebra A over a commutative ring K is a ring R together with a choice of subring K
from its centre.
(2.14.9) If M is a monoid then an algebra A over K is said to be M -graded if
A = mM Am
is a direct sum decomposition of A as a K-module; and
Am Am Amm
(When the latter is replaced by an equality A is strongly graded.)
(2.14.10) Note that if A is a K-algebra, and hence a ring, then A G as above is a strongly
G-graded K-algebra. In particular
(A G)g = Ag