An Introduction to Galois Theory
Andrew Baker
[09/01/2010]
School of Mathematics & Statistics, University of Glasgow.
Email address: a.baker@maths.gla.ac.uk
URL: http://www.maths.gla.ac.uk/∼ajb
Q(
3
√
2, ζ
3
)
Q(
3
√
2)
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
Q(
3
√
2 ζ
3
)
2
.
.
.
.
.
.
.
.
.
.
.
.
3
Q(
3
√
2 ζ
2
3
)
2
3
Q(ζ
3
)
3
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Q
Gal(E/Q)
∼
= S
3
{id, (1 2 3), (1 3 2)}
3
2
{id, (2 3)}
2
3
{id, (1 3)}
2
3
{id, (1 2)}
2
3
{id}
The Galois Correspondence for Q(
3
√
2, ζ
3
)/Q
Introduction: What is Galois Theory?
Much of early algebra centred around the search for explicit formulae for roots of polynomial
equations in one or more unknowns. The solution of linear and quadratic equations in a single
unknown was well understood in antiquity, while formulae for the roots of general real cubics
and quartics was solved by the 16th century. These solutions involved complex numbers rather
than just real numbers. By the early 19th century no general solution of a general polynomial
equation ‘by radicals’ (i.e., by repeatedly taking nth roots for various n) was found despite
considerable eﬀort by many outstanding mathematicians. Eventually, the work of Abel and
Galois led to a satisfactory framework for fully understanding this problem and the realization
that the general polynomial equation of degree at least 5 could not always be solved by radi
cals. At a more profound level, the algebraic structure of Galois extensions is mirrored in the
subgroups of their Galois groups, which allows the application of group theoretic ideas to the
study of ﬁelds. This Galois Correspondence is a powerful idea which can be generalized to apply
to such diverse topics as ring theory, algebraic number theory, algebraic geometry, diﬀerential
equations and algebraic topology. Because of this, Galois theory in its many manifestations is
a central topic in modern mathematics.
In this course we will focus on the following topics.
• The solution of polynomial equations over a ﬁeld, including relationships between roots,
methods of solutions and location of roots.
• The structure of ﬁnite and algebraic extensions of ﬁelds and their automorphisms.
We will study these in detail, building up a theory of algebraic extensions of ﬁelds and their
automorphism groups and applying it to solve questions about roots of polynomial equations.
The techniques we will meet can also be applied to study the following some of which may be
met by people studying more advanced courses.
• Classic topics such as squaring the circle, duplication of the cube, constructible numbers
and constructible polygons.
• Applications of Galois theoretic ideas in Number Theory, the study of diﬀerential
equations and Algebraic Geometry.
There are many good introductory books on Galois Theory, some of which are listed in the
Bibliography. In particular, [2, 3, 8] are all excellent sources and have many similarities to the
present approach to the material.
c ⃝A. J. Baker (2009)
ii
Contents
Introduction: What is Galois Theory? ii
Chapter 1. Integral domains, ﬁelds and polynomial rings 1
Basic notions, convention, etc 1
1.1. Recollections on integral domains and ﬁelds 1
1.2. Polynomial rings 6
1.3. Identifying irreducible polynomials 12
1.4. Finding roots of complex polynomials of small degree 16
1.5. Automorphisms of rings and ﬁelds 19
Exercises on Chapter 1 23
Chapter 2. Fields and their extensions 27
2.1. Fields and subﬁelds 27
2.2. Simple and ﬁnitely generated extensions 29
Exercises on Chapter 2 33
Chapter 3. Algebraic extensions of ﬁelds 35
3.1. Algebraic extensions 35
3.2. Splitting ﬁelds and Kronecker’s Theorem 39
3.3. Monomorphisms between extensions 42
3.4. Algebraic closures 45
3.5. Multiplicity of roots and separability 48
3.6. The Primitive Element Theorem 52
3.7. Normal extensions and splitting ﬁelds 54
Exercises on Chapter 3 55
Chapter 4. Galois extensions and the Galois Correspondence 57
4.1. Galois extensions 57
4.2. Working with Galois groups 58
4.3. Subgroups of Galois groups and their ﬁxed ﬁelds 60
4.4. Subﬁelds of Galois extensions and relative Galois groups 61
4.5. The Galois Correspondence and the Main Theorem of Galois Theory 62
4.6. Galois extensions inside the complex numbers and complex conjugation 64
4.7. Galois groups of even and odd permutations 65
4.8. Kaplansky’s Theorem 68
Exercises on Chapter 4 71
Chapter 5. Galois extensions for ﬁelds of positive characteristic 73
5.1. Finite ﬁelds 73
iii
5.2. Galois groups of ﬁnite ﬁelds and Frobenius mappings 77
5.3. The trace and norm mappings 79
Exercises on Chapter 5 80
Chapter 6. A Galois Miscellany 83
6.1. A proof of the Fundamental Theorem of Algebra 83
6.2. Cyclotomic extensions 84
6.3. Artin’s Theorem on linear independence of characters 88
6.4. Simple radical extensions 90
6.5. Solvability and radical extensions 92
6.6. Symmetric functions 96
Exercises on Chapter 6 97
Bibliography 101
Solutions 103
Chapter 1 103
Chapter 2 110
Chapter 3 112
Chapter 4 114
Chapter 5 117
Chapter 6 119
iv
CHAPTER 1
Integral domains, ﬁelds and polynomial rings
Basic notions, convention, etc
In these notes, a ring will always be a ring with unity 1 ̸= 0. Most of the rings encountered
will also be commutative. An ideal I ▹ R will always mean a twosided ideal. An ideal I ▹ R in
a ring R is proper if I ̸= R, or equivalently if I R. Under a ring homomorphism φ: R −→ S,
1 ∈ R is sent to 1 ∈ S, i.e., φ(1) = 1.
1.1. Definition. Let φ: R −→ S be a ring homomorphism.
• φ is a monomorphism if it is injective, i.e., if for r
1
, r
2
∈ R,
φ(r
1
) = φ(r
2
) =⇒ r
1
= r
2
,
or equivalently if ker φ = {0}.
• φ is an epimorphism if it is surjective, i.e., if for every s ∈ S there is an r ∈ R with
φ(r) = s.
• φ is an isomorphism if it is both a monomorphism and an epimorphism, i.e., if it is
invertible (in which case its inverse is also an isomorphism).
1.1. Recollections on integral domains and ﬁelds
The material in this section is standard and most of it should be familiar. Details may be
found in [3, 5] or other books containing introductory ring theory.
1.2. Definition. A commutative ring R in which there are no zerodivisors is called an
integral domain or an entire ring. This means that for u, v ∈ R,
uv = 0 =⇒ u = 0 or v = 0.
1.3. Example. The following rings are integral domains.
(i) The ring of integers, Z.
(ii) If p is a prime, the ring of integers modulo p, F
p
= Z/p = Z/(p).
(iii) The rings of rational numbers, Q, real numbers, R, and complex numbers, C.
(iv) The polynomial ring R[X], where R is an integral domain; in particular, the polynomial
rings Z[X], Q[X], R[X] and C[X] are all integral domains.
1.4. Definition. Let I ▹ R be a proper ideal in a commutative ring R.
• I is a prime ideal if for u, v ∈ R,
uv ∈ I =⇒ u ∈ I or v ∈ I.
• I is a maximal ideal R if whenever J ▹ R is a proper ideal and I ⊆ J then J = I.
1
• I ▹ R is principal if
I = (p) = {rp : r ∈ R}
for some p ∈ R. Notice that if p, q ∈ R, then (q) = (p) if and only if q = up for some
unit u ∈ R. We also write p  x if x ∈ (p).
• p ∈ R is prime if (p) ▹ R is a prime ideal; this is equivalent to the requirement that
whenever p  xy with x, y ∈ R then p  x or p  y.
• R is a principal ideal domain if it is an integral domain and every ideal I ▹R is principal.
1.5. Example. Every ideal I ▹Z is principal, so I = (n) for some n ∈ Z which we can always
take to be nonnegative, i.e., n 0. Hence Z is a principal ideal domain.
1.6. Proposition. Let R be a commutative ring and I ▹ R an ideal.
(i) The quotient ring R/I is an integral domain if and only if I is a prime ideal.
(ii) The quotient ring R/I is a ﬁeld if and only if I is a maximal ideal.
1.7. Example. If n 0, the quotient ring Z/n = Z/(n) is an integral domain if and only if
n is a prime.
For any (not necessarily commutative) ring with unity there is an important ring homomor
phism η: Z −→ R called the unit or characteristic homomorphism which is deﬁned by
η(n) = n1 =
1 +· · · + 1
. .. .
n
if n > 0,
−(1 +· · · + 1
. .. .
−n
) if n < 0,
0 if n = 0.
Since 1 ∈ R is nonzero, ker η ▹ Z is a proper ideal and using the Isomorphism Theorems we
see that there is a quotient monomorphism η: Z/ ker η −→ R which allows us to identify the
quotient ring Z/ ker η with the image ηZ ⊆ R as a subring of R. By Example 1.5, there is a
unique nonnegative integer p 0 such that ker η = (p); this p is called the characteristic of R
and denoted char R.
1.8. Lemma. If R is an integral domain, its characteristic char R is a prime.
Proof. Consider p = char R. If p = 0 we are done. So suppose that p > 0. The quotient
monomorphism η: Z/ ker η −→ R identiﬁes Z/ ker η with the subring imη = imη of the integral
domain R. But every subring of an integral domain is itself an integral domain, hence Z/ ker η is
an integral domain. Now by Proposition 1.6(i), ker η = (p) is prime ideal and so by Example 1.7,
p is a prime.
1.9. Remark. When discussing a ring with unit R, we can consider it as containing as a
subring of the form Z/(char R) since the quotient homomorphism η: Z/(char R) −→ R gives
an isomorphism Z/(char R) −→ imη, allowing us to identify these rings. In particular, every
integral domain contains as a subring either Z = Z/(0) (if char R = 0) or Z/(p) if p = char R > 0
is a nonzero prime. This subring is sometimes called the characteristic subring of R. The rings
Z and Z/n = Z/(n) for n > 0 are often called core rings. When considering integral domains,
the rings Z and F
p
= Z/p = Z/(p) for p > 0 a prime are called prime rings.
2
Here is a useful and important fact about rings which contain a ﬁnite prime ring F
p
.
1.10. Theorem (Idiot’s Binomial Theorem). Let R be a commutative ring with unit con
taining F
p
for some prime p > 0. If u, v ∈ R, then
(u +v)
p
= u
p
+v
p
.
Proof. We have p1 = 0 in R, hence pt = 0 for any t ∈ R. The Binomial Expansion yields
(1.1) (u +v)
p
= u
p
+
p
1
u
p−1
v +
p
2
u
p−2
v
2
+· · · +
p
p −1
uv
p−1
+v
p
.
Now suppose that 1 j p −1. Then we have
p
j
=
p (p −1)!
j! (p −j)!
= p ×
(p −1)!
j! (p −j)!
.
There are no factors of p appearing in (p −1)!, j! or (p −j)!, so since this number is an integer
it must be divisible by p, i.e.,
(1.2a) p 
p
j
,
or equivalently
(1.2b)
p
j
≡ 0 (mod p).
Hence in R we have
p
j
1 = 0.
Combining the divisibility conditions of (1.2) with the expansion of (1.1), we obtain the
required equation in R,
(u +v)
p
= u
p
+v
p
.
1.11. Definition. Let R be a ring. An element u ∈ R is a unit if it is invertible, i.e., there
is and element v ∈ R for which
uv = 1 = vu.
We usually write u
−1
for this element v, which is necessarily unique and is called the (multi
plicative) inverse of u in R. We will denote the set of all invertible elements of R by R
×
and
note that it always forms a group under multiplication.
1.12. Definition. A commutative ring k is a ﬁeld if every nonzero element u ∈ k is a unit.
This is equivalent to requiring that k
×
= k −{0}.
The familiar rings Q, R and C are all ﬁelds.
1.13. Example. If n 1, the quotient ring Z/n is a ﬁeld if and only if n is a prime.
1.14. Proposition. Every ﬁeld is an integral domain.
Proof. Let k be a ﬁeld. Suppose that u, v ∈ k and uv = 0. If u ̸= 0, we can multiply by
u
−1
to obtain
v = u
−1
uv = 0,
hence v = 0. So at least one of u, v must be 0.
3
A nonzero element p ∈ R is irreducible if for u, v ∈ R,
p = uv =⇒ u or v is a unit.
1.15. Lemma. Let R be an integral domain. If p ∈ R is a nonzero prime then it is an
irreducible.
Proof. Suppose that p = uv for some u, v ∈ R. Then p  u or p  v, and we might as well
assume that u = tp for some t ∈ R. Then (1 −tv)p = 0 and so tv = 1, showing that v is a unit
with inverse t.
Now let D be an integral domain. A natural question to ask is whether D is isomorphic to
a subring of a ﬁeld. This is certainly true for the integers Z which are contained in the ﬁeld of
rational numbers Q, and for a prime p > 0, the prime ring F
p
is itself a ﬁeld.
1.16. Definition. The ﬁelds Q and F
p
where p > 0 is a prime are the prime ﬁelds.
Of course, we can view Z as a subring of any subﬁeld of the complex numbers so an answer
to this question may not be unique! However, there is always a ‘smallest’ such ﬁeld which is
unique up to an isomorphism.
1.17. Theorem. Let D be an integral domain.
(i) There is a ﬁeld of fractions of D, Fr(D), which contains D as a subring.
(ii) If φ: D −→ F is a ring monomorphism into a ﬁeld F, there is a unique homomorphism
¯ φ: Fr(D) −→ F such that ¯ φ(t) = φ(t) for all t ∈ D ⊆ Fr(D).
D
inc
φ
F
Fr(D)
∃! ¯ φ
Proof. (i) Consider the set
P(D) = {(a, b) : a, b ∈ D, b ̸= 0}.
Now introduce an equivalence relation ∼ on P(D), namely
(a
′
, b
′
) ∼ (a, b) ⇐⇒ ab
′
= a
′
b.
Of course, it is necessary to check that this relation is an equivalence relation; this is left as an
exercise. We denote the equivalence class of (a, b) by [a, b] and the set of equivalence classes by
Fr(D).
We deﬁne addition and multiplication on Fr(D) by
[a, b] + [c, d] = [ad +bc, bd], [a, b][c, d] = [ac, bd].
We need to verify that these operations are well deﬁned. For example, if [a
′
, b
′
] = [a, b] and
[c
′
, d
′
] = [c, d], then
(a
′
d
′
+b
′
c
′
)bd = a
′
d
′
bd +b
′
c
′
bd = ab
′
d
′
d +b
′
bcd
′
= (ad +bc)b
′
d
′
,
and so (a
′
d
′
+ b
′
c
′
, b
′
d
′
) ∼ (ad + bc, bd); hence addition is well deﬁned. A similar calculation
shows that (a
′
c
′
, b
′
d
′
) ∼ (ac, bd), so multiplication is also well deﬁned. It is now straightforward
4
to show that Fr(D) is a commutative ring with zero 0 = [0, 1] and unit 1 = [1, 1]. In fact, as we
will soon see, Fr(D) is a ﬁeld.
Let [a, b] ∈ Fr(D). Then [a, b] = [0, 1] if and only if (0, 1) ∼ (a, b) which is equivalent to
requiring that a = 0; notice that for any b ̸= 0, [0, b] = [0, 1]. We also have [a, b] = [1, 1] if and
only if a = b.
Now let [a, b] ∈ Fr(D) be nonzero, i.e., a ̸= 0. Then b ̸= 0 and [a, b], [b, a] ∈ Fr(D) satisfy
[a, b][b, a] = [ab, ba] = [1, 1] = 1,
so [a, b] has [b, a] as an inverse. This shows that Fr(D) is a ﬁeld.
We can view D as a subring of Fr(D) using the map
j : D −→ Fr(D); j(t) = [t, 1]
which is a ring homomorphism; it is easy to check that it is a monomorphism. Therefore we
may identify t ∈ D with j(t) = [t, 1] ∈ Fr(D) and D with the subring imj ⊆ Fr(D).
(ii) Consider the function
Φ: P(D) −→ F; Φ(a, b) = φ(a)φ(b)
−1
.
If (a
′
, b
′
) ∼ (a, b), then
Φ(a
′
, b
′
) = φ(a
′
)φ(b
′
)
−1
= φ(a
′
)φ(b)φ(b)
−1
φ(b
′
)
−1
= φ(a
′
b)φ(b)
−1
φ(b
′
)
−1
= φ(ab
′
)φ(b
′
)
−1
φ(b)
−1
= φ(a)φ(b
′
)φ(b
′
)
−1
φ(b)
−1
= φ(a)φ(b)
−1
= Φ(a, b),
so Φ is constant on each equivalence class of ∼. Hence we may deﬁne the function
¯ φ: Fr(D) −→ F; ¯ φ([a, b]) = Φ(a, b).
It is now easy to verify that ¯ φ is a ring homomorphism which agrees with φ on D ⊆ Fr(D).
The next three corollaries are left as an exercise.
1.18. Corollary. If F is a ﬁeld then F = Fr(F).
1.19. Corollary. If D is a subring of a ﬁeld F, then Fr(D) ⊆ Fr(F) = F and Fr(D) is
the smallest subﬁeld of F containing D.
1.20. Corollary. Let D
1
and D
2
be integral domains and let φ: D
1
−→ D
2
be a ring
monomorphism. Then there is a unique induced ring homomorphism φ
∗
: Fr(D
1
) −→ Fr(D
2
)
which satisﬁes φ
∗
(t) = φ(t) whenever t ∈ D
1
⊆ Fr(D
1
).
D
1
φ
inc
D
2
inc
Fr(D
1
)
φ
∗
Fr(D
2
)
Moreover, this construction has the following properties.
5
• If φ: D
1
−→ D
2
and θ: D
2
−→ D
3
are monomorphisms between integral domains then
θ
∗
◦ φ
∗
= (θ ◦ φ)
∗
as homomorphisms Fr(D
1
) −→ Fr(D
3
).
• For any integral domain D, the identity homomorphism id: D −→ D induces the
identity homomorphism (id)
∗
= id: Fr(D) −→ Fr(D).
D
1
φ
inc
D
2
θ
inc
D
3
inc
Fr(D
1
)
φ
∗
Fr(D
2
)
θ
∗
Fr(D
3
)
D
id
inc
D
inc
Fr(D)
id
∗
= id
Fr(D)
1.21. Remarks. (a) When working with a ﬁeld of fractions we usually adopt the familiar
notation
a
b
= a/b = [a, b]
for the equivalence class of (a, b). The rules for algebraic manipulation of such symbols are the
usual ones for working with fractions, i.e.,
a
1
b
1
+
a
2
b
2
=
a
1
b
2
+a
2
b
1
b
1
b
2
,
a
1
b
1
×
a
2
b
2
=
a
1
b
1
a
2
b
2
=
a
1
a
2
b
1
b
2
.
The ﬁeld of fractions of an integral domain is sometimes called its ﬁeld of quotients, however as
the word quotient is also associated with quotient rings we prefer to avoid using that terminology.
(b) Corollary 1.20 is sometimes said to imply that the construction of Fr(D) is functorial in the
integral domain D.
1.2. Polynomial rings
Let R be a commutative ring. We will make frequent use of the ring R[X] of polynomials
over R in an indeterminate X. This consists of elements of form
p(X) = p
0
+p
1
X +· · · +p
m
X
m
where m 0 and p
0
, p
1
, . . . , p
m
∈ R; such p(X) are called polynomials. Addition and multipli
cation in R[X] are deﬁned by
(p
0
+p
1
X +· · · +p
m
X
m
) + (q
0
+q
1
X +· · · +q
m
X
m
) =
(p
0
+q
0
) + (p
1
+q
1
)X +· · · + (p
m
+q
m
)X
m
,
and
(p
0
+p
1
X +· · · +p
m
X
m
)(q
0
+q
1
X +· · · +q
m
X
m
) =
(p
0
q
0
) + (p
0
q
1
+p
1
q
0
)X +· · · + (p
0
q
m
+p
1
q
m−1
+· · · +p
m−1
q
1
+p
m
q
0
)X
2m
.
Then R[X] is a commutative ring with the constant polynomials 0 and 1 as its zero and unit.
We identify r ∈ R with the obvious constant polynomial; this allows us to view R as a subring
of R[X] and the inclusion function inc: R −→ R[X] is a monomorphism.
More generally, we inductively can deﬁne the ring of polynomials in n indeterminates
X
1
, . . . , X
n
over R,
R[X
1
, . . . , X
n
] = R[X
1
, . . . , X
n−1
][X
n
]
6
for n 1. Again there is an inclusion monomorphism inc: R −→ R[X
1
, . . . , X
n
] which sends
each element of R to itself considered as a constant polynomial.
These polynomial rings have an important universal property.
1.22. Theorem (Homomorphism Extension Property). Let φ: R −→ S be a ring homo
morphism.
(i) For each s ∈ S there is a unique ring homomorphism φ
s
: R[X] −→ S for which
• φ
s
(r) = φ(r) for all r ∈ R,
• φ
s
(X) = s.
R
inc
φ
S
R[X]
∃! φ
s
(ii) For n 1 and s
1
, . . . , s
n
∈ S, there is a unique ring homomorphism
φ
s
1
,...,s
n
: R[X
1
, . . . , X
n
] −→ S
for which
• φ
s
1
,...,s
n
(r) = φ(r) for all r ∈ R,
• φ
s
1
,...,s
n
(X
i
) = s
i
for i = 1, . . . , n.
R
inc
φ
S
R[X
1
, . . . , X
n
]
∃! φ
s
1
,...,s
n
Proof. (Sketch)
(i) For a polynomial p(X) = p
0
+p
1
X +· · · +p
m
X
m
∈ R[X], we deﬁne
(1.3) φ
s
(p(X)) = p
0
+p
1
s +· · · +p
m
s
m
∈ S.
It is then straightforward to check that φ
s
is a ring homomorphism with the stated properties
and moreover is the unique such homomorphism.
(ii) is proved by induction on n using (i).
We will refer to φ
s
1
,...,s
n
as the extension of φ by evaluation at s
1
, . . . , s
n
. It is standard to
write
p(s
1
, . . . , s
n
) = φ
s
1
,...,s
n
(p(X
1
, . . . , X
n
)).
An extremely important special case occurs when we start with the identity homomorphism
id: R −→ R and r
1
, . . . , r
n
∈ R; then we have the homomorphism
ε
r
1
,...,r
n
= id
r
1
,...,r
n
: R[X
1
, . . . , X
n
] −→ R.
Slightly more generally we may take the inclusion of a subring inc: R −→ S and s
1
, . . . , s
n
∈ S;
then
ε
s
1
,...,s
n
= inc
s
1
,...,s
n
: R[X
1
, . . . , X
n
] −→ S
is called evaluation at s
1
, . . . , s
n
and we denote its image by
R[s
1
, . . . , s
n
] = ε
s
1
,...,s
n
R[X
1
, . . . , X
n
] ⊆ S.
7
Then R[s
1
, . . . , s
n
] is a subring of S, called the subring generated by s
1
, . . . , s
n
over R.
Here is an example illustrating how we will use such evaluation homomorphisms.
1.23. Example. Consider the inclusion homomorphism inc: Q −→ C. We have the evalu
ation at i homomorphism ε
i
, for which ε
i
(X) = i. We easily see that ε
i
Q[X] ⊆ C is a subring
Q[i] ⊆ C consisting of the complex numbers of form a +bi with a, b ∈ Q.
Notice that if we had used −i instead of i, evaluation at −i, ε
−i
, we would also have
ε
−i
Q[X] = Q[i]. These evaluation homomorphisms are related by complex conjugation since
ε
−i
(p(X)) = ε
i
(p(X)),
which is equivalent to the functional equation
ε
−i
= ( ) ◦ ε
i
.
Notice also that in these examples we have
ker ε
−i
= ker ε
i
= (X
2
+ 1) ▹ Q[X],
hence we also have
Q[i]
∼
= Q[X]/(X
2
+ 1).
In fact (X
2
+ 1) is actually a maximal ideal and so Q[i] ⊆ C is a subﬁeld; later we will write
Q(i) for this subﬁeld.
1.24. Proposition. Let R be an integral domain.
(i) The ring R[X] of polynomials in an indeterminate X over R is an integral domain.
(ii) The ring R[X
1
, . . . , X
n
] of polynomials in the indeterminates X
1
, . . . , X
n
over R is an
integral domain.
1.25. Corollary. Let k be a ﬁeld and n 1. Then the polynomial ring k[X
1
, . . . , X
n
] in
the indeterminates X
1
, . . . , X
n
is an integral domain.
As we will make considerable use of such rings we describe in detail some of their important
properties. First we recall long division in a polynomial ring k[X] over a ﬁeld k; full details can
be found in a basic course on commutative rings or any introductory book on this subject.
1.26. Theorem (Long Division). Let k be a ﬁeld. Let f(X), d(X) ∈ k[X] and assume that
d(X) ̸= 0 so that deg d(X) > 0. Then there are unique polynomials q(X), r(X) ∈ k[X] for
which
f(X) = q(X)d(X) +r(X)
and either deg r(X) < deg d(X) or r(X) = 0.
In the situation discussed in this result, the following names are often used. We refer to the
process of ﬁnding q(X) and r(X) as long division of f(X) by d(X). Also,
f(X) = the dividend, d(X) = the divisor, q(X) = the quotient, r(X) = the remainder.
1.27. Example. For k = Q, ﬁnd the quotient and remainder when f(X) = 6X
4
− 6X
3
+
3X
2
−3X + 1 is divided by d(X) = 2X
2
+ 1.
8
Solution. In the usual notation we have the following calculation.
3X
2
−3X
2X
2
+ 1  6X
4
−6X
3
+ 3X
2
−3X + 1
6X
4
+ 0X
3
+ 3X
2
+ 0X + 0
−6X
3
+ 0X
2
−3X + 1
−6X
3
+ 0X
2
−3X + 0
1
Hence
6X
4
−6X
3
+ 3X
2
−3X + 1 = (3X
2
−3X)(2X
2
+ 1) + 1,
giving q(X) = 3X
2
−3X and r(X) = 1.
1.28. Example. For k = F
5
, ﬁnd the quotient and remainder when f(X) = 10X
5
+ 6X
4
−
6X
3
+ 3X
2
−3X + 1 is divided by d(X) = 2X
2
+ 1.
Solution. First notice that working modulo 5 we have
f(X) = 10X
5
+ 6X
4
−6X
3
+ 3X
2
−3X + 1 ≡ X
4
+ 4X
3
+ 3X
2
+ 2X + 1 (mod 5).
Notice also following multiplicative inverses in F
5
:
2
−1
≡ 3 (mod 5), 3
−1
≡ 2 (mod 5), 4
−1
≡ 4 (mod 5).
We have the following calculation.
3X
2
+ 2X
2X
2
+ 1 6X
4
+ 4X
3
+ 3X
2
+ 2X + 1
6X
4
+ 0X
3
+ 3X
2
+ 0X + 0
4X
3
+ 0X
2
+ 2X + 1
4X
3
+ 0X
2
+ 2X + 0
1
Hence
6X
4
−6X
3
+ 3X
2
−3X + 1 ≡ (3X
2
+ 2X)(2X
2
+ 1) + 1 (mod 5),
giving q(X) = 3X
2
+ 2X and r(X) = 1.
An important consequence of Theorem 1.26 is the following which makes use of the Euclidean
Algorithm.
1.29. Corollary. Let k be a ﬁeld and X an indeterminate. Let f(X), g(X) ∈ k[X] be
nonzero. Then there are a(X), b(X) ∈ k[X] such that
a(X)f(X) +b(X)g(X) = gcd(f(X), g(X)).
Here the greatest common divisor gcd(f(X), g(X)) of f(X), g(X) is the monic polynomial
of greatest degree which divides both of f(X), g(X).
9
1.30. Proposition. Let k be a ﬁeld and X an indeterminate. Then a nonconstant poly
nomial p(X) ∈ k[X] is an irreducible if and only if it is a prime.
Proof. By Lemma 1.15 we already know that p(X) is irreducible if it is prime. So sup
pose that p(X) is irreducible and that p(X)  u(X)v(X) for u(X), v(X) ∈ k[X]. Then by
Corollary 1.29, there are a(X), b(X) ∈ k[X] such that
a(X)p(X) +b(X)u(X) = gcd(p(X), u(X)).
But since p(X) is irreducible, gcd(p(X), u(X)) = p(X) or gcd(p(X), u(X)) = 1. In the latter
case,
a(X)p(X) +b(X)u(X) = 1,
and multiplying through by v(X) gives
a(X)p(X)v(X) +b(X)u(X)v(X) = v(X)
and so p(X)  v(X). This shows that p(X)  u(X) or p(X)  v(X), and so p(X) is prime.
1.31. Theorem. Let k be a ﬁeld and X an indeterminate.
(i) Every ideal I ▹ k[X] is principal, i.e., I = (h(X)) for some h(X) ∈ k[X].
(ii) The ideal (p(X)) ▹k[X] is prime if and only if p(X) = 0 or p(X) is irreducible in k[X].
(iii) The quotient ring k[X]/(p(X)) is an integral domain if and only if p(X) = 0 or p(X)
is irreducible in k[X].
(iv) The quotient ring k[X]/(p(X)) is a ﬁeld if and only if p(X) is an irreducible in k[X].
Proof. (i) Let I ▹ k[X] and assume that I ̸= (0). Then there must be at least one element
of I with positive degree and so we can choose h(X) ∈ I of minimal degree, say d = deg h(X).
Now let p(X) ∈ I. By Long Division, there are q(X), r(X) ∈ k[X] such that
p(X) = q(X)h(X) +r(X) and deg r(X) < d or r(X) = 0.
Since p(X) and h(X) are in the ideal I, we also have
r(X) = p(X) −q(X)h(X) ∈ I.
If r(X) ̸= 0, this would contradict the minimality of d, so we must have r(X) = 0, showing that
p(X) = q(X)h(X). Thus I ⊆ (p(X)) ⊆ I and therefore I = (p(X)).
(ii) This follows from Proposition 1.30.
(iii) This follows from Proposition 1.6(i).
(iv) Since k[X] is an integral domain and not a ﬁeld, it follows that if k[X]/(p(X)) is a ﬁeld
then because it is an integral domain, p(X) is an irreducible by (iii).
Suppose that p(X) is irreducible (and hence is nonzero). Then for any q(X) ∈ k[X] with
q(X) / ∈ (p(X)), by Corollary 1.29 we can ﬁnd suitable a(X), b(X) ∈ k[X] for which
a(X)p(X) +b(X)q(X) = gcd(p(X), q(X)).
But gcd(p(X), q(X)) = 1 since p(X) is irreducible, so
a(X)p(X) +b(X)q(X) = 1.
This shows that in the quotient ring k[X]/(p(X)) the residue class of q(X) has the residue class
of b(X) as its inverse.
10
1.32. Remark. In connection with Theorem 1.31(i), notice that if p(X) ∈ k[X], then
provided d = deg p(X) > 0, we have for some p
d
̸= 0,
p(X) = p
0
+p
1
X +· · · +p
d
X
d
= p
d
q(X),
where
q(X) = p
−1
d
p
0
+p
−1
d
p
1
X +· · · +p
−1
d
p
d−1
X
d−1
+X
d
.
This easily implies that as ideals of k[X], (p(X)) = (q(X)). So we can always ﬁnd a monic
polynomial as the generator of a given ideal, and this monic polynomial is unique.
1.33. Proposition (Unique Factorization Property). Every nonconstant polynomial f(x) ∈
k[X] has a factorization
f(x) = cp
1
(X) · · · p
k
(X),
where c ∈ k, and p
1
(X), . . . , p
k
(X) ∈ k[X] are irreducible monic polynomials. Moreover, c is
unique and the sequence of polynomials p
1
(X), . . . , p
k
(X) is unique apart from the order of the
terms.
Proof. (Sketch)
Existence is proved by induction on the degree of f(X) and begins with the obvious case
deg f(X) = 1. If deg f(X) > 1, then either f(X) is already irreducible, or f(X) = f
1
(X)f
2
(X)
with both factors of positive degree, and therefore deg f
j
(X) < deg f(X). This gives the
inductive step.
To prove uniqueness, suppose that
p
1
(X) · · · p
k
(X) = q
1
(X) · · · q
ℓ
(X)
where p
i
(X), q
j
(X) ∈ k[X] are irreducible monic polynomials. Then by Proposition 1.30, each
p
i
(X) is prime, hence divides one of the q
j
(X), hence must equal it. By reordering we can
assume that p
i
(X) = q
i
(X) and k ℓ. After cancelling common factors we obtain
q
k+1
(X) · · · q
ℓ
(X) = 1,
and so we see that k = ℓ.
1.34. Corollary. Suppose that f(X) ∈ k[X] factors into linear factors
f(X) = c(X −u
1
) · · · (X −u
d
),
where u
1
, . . . , u
d
∈ k. Then the sequence of roots u
1
, . . . , u
d
is unique apart from the order. In
particular, if v
1
, . . . , v
r
are the distinct roots, then
f(X) = c(X −v
1
)
m
1
· · · (X −v
r
)
m
r
,
where m
i
> 0 and this factorization is unique apart from the order of the pairs (v
i
, m
i
).
1.35. Corollary. The number of distinct roots of a nonconstant polynomial f(X) ∈ k[X]
is at most deg f(X).
1.36. Definition. If k is a ﬁeld and X an indeterminate, then the ﬁeld of fractions of k[X]
is the ﬁeld of rational functions, k(X). The elements of k(X) are fractions of the form
a
0
+a
1
X +· · · +a
m
X
m
b
0
+b
1
X +· · · +b
n
X
n
with a
i
, b
j
∈ k and b
0
+b
1
X +· · · +b
n
X
n
̸= 0.
11
1.3. Identifying irreducible polynomials
When k is a ﬁeld, we will need some eﬀective methods for deciding when a polynomial in
k[X] is irreducible.
Let us consider factorisation of polynomials over Q. If f(X) ∈ Z[X] then we can also consider
f(X) as an element of Q[X]. If R = Z or Q, we say that f(X) has a proper factorisation over
R if f(X) = g(X)h(X) for some g(X), h(X) ∈ R[X] with deg g(X) > 0 and deg h(X) > 0.
1.37. Proposition (Gauss’s Lemma). Let f(X) ∈ Z[X]. Then f(X) has a proper factori
sation over Z if and only it has a proper factorisation over Q.
So to ﬁnd factors of f(X) it is suﬃcient to look for factors in Z[X]. Our next result is a
special case of the Eisenstein Irreducibility Test. The version here is slightly more general than
the more usual one which corresponds to taking s = 0.
1.38. Proposition (Eisenstein Test). Let f(X) ∈ Z[X] and s ∈ Z. Choose a
i
∈ Z so that
f(X) = a
0
+a
1
(X −s) +· · · +a
d−1
(X −s)
d−1
+a
d
(X −s)
d
,
where d = deg f(X). Suppose that p > 0 is a prime for which the following three conditions
hold:
• a
k
≡ 0 (mod p) for k = 0, . . . , d −1;
• a
0
̸≡ 0 (mod p
2
);
• a
d
̸≡ 0 (mod p).
Then f(X) is irreducible in Q[X] and hence also in Z[X].
1.39. Example. Let p 2 be a prime. Then the polynomial
Φ
p
(X) = 1 +X +· · · +X
p−1
∈ Z[X]
is irreducible in Q[X] and hence also in Z[X].
Proof. Working in Z[X],
Φ
p
(X)(X −1) = (1 +X +· · · +X
p−1
)(X −1)
= X
p
−1
= (1 + (X −1))
p
−1
=
p
∑
k=1
p
k
(X −1)
k
≡ (X −1)
p
(mod p),
since by (1.2a), p divides
p
k
=
p!
k! (p −k)!
when k = 1, . . . , p −1. Hence
Φ
p
(X) ≡ (X −1)
p−1
(mod p)
Also,
p
1
= p ̸≡ 0 (mod p
2
),
12
giving
(1.4) Φ
p
(X) = (X −1)
p−1
+c
p−2
(X −1)
p−2
+· · · +c
1
(X −1) +c
0
with c
r
≡ 0 (mod p) and c
0
= p. So the Eisenstein Test can be applied here with s = 1 to show
that Φ
p
(X) is irreducible in Z[X].
1.40. Example. As examples we have the irreducible polynomials
Φ
2
(X) = 1 +X,
Φ
3
(X) = 1 +X +X
2
,
Φ
5
(X) = 1 +X +X
2
+X
3
+X
4
,
Φ
7
(X) = 1 +X +X
2
+X
3
+X
4
+X
5
+X
6
,
Φ
11
(X) = 1 +X +X
2
+X
3
+X
4
+X
5
+X
6
+X
7
+X
8
+X
9
+X
10
.
These are examples of the cyclotomic polynomials Φ
n
(X) ∈ Z[X] which are deﬁned for all
n 1 by
(1.5a) X
n
−1 =
∏
dn
Φ
d
(X),
where the product is taken over all the positive divisors of n. For example,
X
2
−1 = (X −1)(X + 1) = Φ
1
(X)Φ
2
(X),
X
3
−1 = (X −1)(X
2
+X + 1) = Φ
1
(X)Φ
3
(X),
X
4
−1 = (X −1)(X + 1)(X
2
+ 1) = Φ
1
(X)Φ
2
(X)Φ
4
(X),
X
5
−1 = (X −1)(X
4
+X
3
+X + 1) = Φ
1
(X)Φ
5
(X),
X
6
−1 = (X −1)(X + 1)(X
2
+X + 1)(X
2
−X + 1) = Φ
1
(X)Φ
2
(X)Φ
3
(X)Φ
6
(X),
X
12
−1 = (X −1)(X + 1)(X
2
+X + 1)(X
2
+ 1)(X
2
−X + 1)(X
4
−X
2
+ 1)
= Φ
1
(X)Φ
2
(X)Φ
3
(X)Φ
4
(X)Φ
6
(X)Φ
12
(X).
Cyclotomic polynomials can be computed recursively using Equation (1.5a). If we know Φ
k
(X)
for k < n, then
(1.5b) Φ
n
(X) =
X
n
−1
∏
dn
d<n
Φ
d
(X)
.
The degree of Φ
n
(X) involves a function of n probably familiar from elementary Number Theory.
1.41. Definition. The Euler function φ: N −→N is deﬁned by
φ(n) = number of k = 1, . . . , n for which gcd(n, k) = 1
= (Z/n)
×
 = number of units in Z/n
= number of generators of the cyclic group Z/n.
In particular, if p 2 is a prime then φ(p) = p −1. Of course, φ(1) = 1.
13
It can be shown that for each natural number n,
(1.6)
∑
dn
φ(d) = n.
Notice that we can inductively determine φ(n) using this equation. For example, if p and q are
distinct primes, then
φ(pq) = pq −(φ(p) +φ(q) +φ(1)) = pq −(p −1) −(q −1) −1 = (p −1)(q −1).
It is also true that whenever m, n are coprime, i.e., when gcd(m, n) = 1,
(1.7) φ(mn) = φ(m)φ(n).
Thus if n = p
r
1
1
· · · p
r
s
s
where p
1
< p
2
< · · · < p
s
are the prime factors of n and r
j
> 0, then
(1.8) φ(n) = φ(p
r
1
1
) · · · φ(p
r
s
s
).
Furthermore, if p is a prime and r > 0, then
(1.9) φ(p
r
) = (p −1)p
r−1
.
Notice that as a result, φ(n) is even when n > 2.
1.42. Remark. For those who know about the M¨obius function µ (which takes values 0, ±1)
and M¨obius inversion, the latter can be used to solve Equation (1.6) for φ, giving
(1.10) φ(n) =
∑
dn
µ(d)
n
d
.
Similarly, the formulae of (1.5) lead to
(1.11) Φ
n
(X) =
∏
dn
(X
n/d
−1)
µ(d)
.
So for example, if p, q are distinct primes, then using standard properties of µ,
Φ
pq
(X) = (X
pq
−1)
µ(1)
(X
pq/p
−1)
µ(p)
(X
pq/q
−1)
µ(q)
(X
pq/pq
−1)
µ(pq)
= (X
pq
−1)(X
q
−1)
−1
(X
p
−1)
−1
(X −1) =
(X
pq
−1)(X −1)
(X
q
−1)(X
p
−1)
.
Recall that an element ζ of a ﬁeld K is a primitive nth root of unity if
min{k : 1 k and ζ
k
= 1} = n.
We think of ζ
n
= e
2πi/n
as the standard complex primitive nth root of unity. Then every
complex nth root of unity has the form ζ
k
n
= e
2πik/n
for k = 0, 1, . . . , n −1.
1.43. Theorem. For each n 1, the cyclotomic polynomial Φ
n
(X) is irreducible in Q[X]
and hence in Z[X]. The complex roots of Φ
n
(X) are the primitive nth roots of unity,
ζ
k
n
= e
2πik/n
(0 k n −1, gcd(k, n) = 1).
and the number of these is deg Φ
n
(X) = φ(n). Hence,
Φ
n
(X) =
∏
t=1,...,n−1
gcd(t,n)=1
(X −ζ
t
n
).
Proof. We will give a reformulation and proof of this in Theorem 6.2.
14
1.44. Example. For n = 6 we have
ζ
6
= e
2πi/6
= e
πi/3
=
1
2
+
√
3
2
i.
Then φ(6) = 2 and
Φ
6
(X) = X
2
−X + 1 = (X −ζ
6
)(X −ζ
5
6
).
It is also worth recording a related general result on cyclic groups.
1.45. Proposition. Let n 1 and C = ⟨g⟩ be a cyclic group of order n and a generator g.
Then an element g
r
∈ C is a generator if and only if gcd(r, n) = 1; the number of such elements
of C is φ(n).
This leads to a useful group theoretic result.
1.46. Lemma. Let G be a ﬁnite group satisfying the following condition:
• For each n 1, there are at most n solutions of x
n
= ι in G.
Then G is cyclic and in particular is abelian.
Proof. Let θ
G
(d) denote the number of elements in G of order d. By Lagrange’s Theorem,
θ
G
(d) = 0 unless d divides G. Since
G =
∪
dG
{g ∈ G : g = d},
we have
G =
∑
dG
θ
G
(d).
Recall the Euler φfunction satisﬁes Equation (1.6), hence
G =
∑
dG
φ(d).
Combining these we obtain
(1.12)
∑
dG
θ
G
(d) =
∑
dG
φ(d).
Let d be a divisor of G. By Proposition 1.45, for each element g ∈ G of order d, the cyclic
subgroup ⟨g⟩ G has φ(d) generators, each of order d. As there are at most d such elements g
in G, this gives θ
G
(d) φ(d). So
∑
dG
θ
G
(d)
∑
dG
φ(d).
Now if θ
G
(d) < φ(d) for some d, we would have a strict inequality in place of Equation (1.12).
Hence θ
G
(d) = φ(d) for all d. In particular, there are φ(G) elements of order G, hence there
must be an element of order G, so G is cyclic.
The above results for polynomials over Q and Z have analogues over the ﬁeld of fractions
k(T) and polynomial ring k[T], where k is a ﬁeld.
A polynomial f(X) ∈ k[T][X] is an element of k(T)[X]. If R = k[T] or k(T), we say that
f(X) has a proper factorisation over R if f(X) = g(X)h(X) for some g(X), h(X) ∈ R[X] with
deg g(X) > 0 and deg h(X) > 0.
15
1.47. Proposition (Gauss’s Lemma). Let f(X) ∈ k[T][X]. Then f(X) has a proper fac
torisation over k[T] if and only it has a proper factorisation over k(T).
Here is another version of the Eisenstein Test; again we state a version which is slightly
more general than the usual one which corresponds to the case where s = 0.
1.48. Proposition (Eisenstein Test). Let f(X) ∈ k[T][X] and s ∈ k[T]. Choose a
i
∈ k[T]
so that
f(X) = a
0
+a
1
(X −s) +· · · +a
d−1
(X −s)
d−1
+a
d
(X −s)
d
,
where d = deg f(X). Suppose that p(T) ∈ k[T] is an irreducible for which the following three
conditions hold:
• a
k
≡ 0 (mod p(T)) for k = 0, . . . , d −1;
• a
0
̸≡ 0 (mod p(T)
2
);
• a
d
̸≡ 0 (mod p(T)).
Then f(X) is irreducible in k(T)[X] and hence also in k[T][X].
1.49. Example. Let k be a ﬁeld. Then the polynomial X
n
−T is irreducible in k(T)[X].
1.4. Finding roots of complex polynomials of small degree
♠
♡ ♢
♣
In this section we work within the complex numbers and take k ⊆ C. In practice we
will usually have k = R or k = C.
For monic linear (degree 1) or quadratic (degree 2) polynomials, methods of ﬁnding roots are
very familiar. Let us consider the cases of cubic (degree 3) and quartic (degree 4) polynomials.
Cubic polynomials: Cardan’s method. The following 16th century method of ﬁnding
roots of cubics is due to Jerˆome Cardan who seems to have obtained some preliminary versions
from Niccol`o Tartaglia by somewhat disreputable means! For historical details see [2, 3].
A monic cubic
f(X) = X
3
+a
2
X
2
+a
1
X +a
0
∈ C[X]
can be transformed into one with no quadratic term by a change of variables X −→ X −a
2
/3
giving
g(X) = f(X −a
2
/3) = X
3
−
a
1
−
1
3
a
2
2
X +
a
0
−
a
1
a
2
3
+
2a
3
2
27
∈ C[X].
Clearly ﬁnding the roots of f(X) is equivalent to ﬁnding those of g(X), so we may as well
assume that we want to ﬁnd the complex roots of
f(X) = X
3
+pX +q ∈ C[X].
Suppose that x ∈ C is a root of f(X), i.e.,
(1.13) x
3
+px +q = 0.
If we introduce u ∈ C for which
x = u −
p
3u
,
then
u −
p
3u
3
+p
u −
p
3u
+q = 0
16
and so
u
3
−
p
3
27u
3
+q = 0,
hence
u
6
+qu
3
−
p
3
27
= 0.
Solving for u
3
we obtain
u
3
= −
q
2
±
1
2
√
q
2
+
4p
3
27
,
where
√
q
2
+
4p
3
27
denotes one of the complex square roots of the discriminant of the quadratic
equation
U
2
+qU −
p
3
27
= 0.
Now if we take u to be a cube root of one of the complex numbers
−
q
2
±
1
2
√
q
2
+
4p
3
27
we obtain the desired root of f(X) as x = u − p/3u. Notice that we have a choice of 2 values
for u
3
and for each of these a choice of 3 values for u, diﬀering by factors of the form ω
r
for
r = 0, 1, 2 where ω = e
2πi/3
is a primitive cube root of 1. However, since
1
−q +
√
q
2
+
4p
3
27
=
−q −
√
q
2
+
4p
3
27
q
2
−(q
2
+ 4p
3
/27)
= −27
−q −
√
q
2
+
4p
3
27
4p
3
,
it is easy to verify that there are in fact only 3 choices of the root x which we can write
symbolically as
x =
3
√
−
q
2
+
1
2
√
q
2
+
4p
3
27
+
3
√
−
q
2
−
1
2
√
q
2
+
4p
3
27
(1.14)
or more precisely as
x =
3
√
−
q
2
+
1
2
√
q
2
+
4p
3
27
−
p
3
3
√
−
q
2
+
1
2
√
q
2
+
4p
3
27
. (1.15)
1.50. Example. Find the complex roots of the polynomial
f(X) = X
3
+ 3X −10 ∈ R[X].
Solution. Applying the method above, we reduce to the quadratic equation
U
2
−10U −1 = 0
whose roots are 5 ±
√
26 ∈ R. Notice that 5 +
√
26 > 0 and 5 −
√
26 < 0; we also have
5 −
√
26 =
−1
5 +
√
26
.
Now 5 +
√
26 has the complex cube roots
3
√
5 +
√
26,
3
√
5 +
√
26 ω,
3
√
5 +
√
26 ω
2
.
17
Here we have x = u −1/u, so the 3 complex roots of f(X) are
3
√
5 +
√
26 −
1
3
√
5 +
√
26
ω
r
(r = 0, 1, 2).
Notice that one of these is real, namely
3
√
5 +
√
26 −
1
3
√
5 +
√
26
=
3
√
5 +
√
26
2
−1
3
√
5 +
√
26
.
Quartic polynomials: Ferrari’s method. The following method of ﬁnding roots of
quartics was publicised by Cardan who attributed it to his student Lodovicio Ferrari.
A general monic quartic polynomial
f(X) = X
4
+a
3
X
3
+a
2
X
2
+a
1
X +a
0
∈ C[X]
can be transformed into one with no cubic term by a change of variables X −→ X−a
2
/3 giving
g(X) = f(X −a
3
/4) =
Y
4
+
a
2
−
3
8
a
2
3
Y
2
+
1
8
a
3
3
−
1
2
a
2
a
3
+a
1
Y −
1
16
a
2
a
2
3
−
3
256
a
4
3
+
1
4
a
1
a
3
+a
0
.
Clearly ﬁnding the roots of f(X) is equivalent to ﬁnding those of g(X), so we may as well
assume that we want to ﬁnd the complex roots of
f(X) = X
4
+pX
2
+qX +r ∈ C[X].
Suppose that x is a root and introduce numbers y, z such that z = x
2
+ y (we will ﬁx the
values of these later). Then
z
2
= x
4
+ 2x
2
y +y
2
= −px
2
−qx −r + 2x
2
y +y
2
= (2y −p)x
2
−qx +y
2
−r.
Now choose y to make the last quadratic expression in x a square,
(1.16) (2y −p)x
2
−qx + (y
2
−r) = (Ax +B)
2
.
This can be done by requiring the vanishing of the discriminant
(1.17) q
2
−4(2y −p)(y
2
−r) = 0.
Notice that if y = p/2 then we would require q = 0 and then
f(X) = X
4
+pX
2
+r = (X
2
)
2
+p(X
2
) +r = 0
can be solved by solving
Z
2
+pZ +r = 0.
Since Equation (1.17) is a cubic in y, we can use the method of solution of cubics to ﬁnd a root
y = t say. Then for Equation (1.16) we have
(x
2
+t)
2
= (Ax +B)
2
,
whence
x
2
= −t ±(Ax +B).
18
Thus taking the two square roots of the right hand side we obtain 4 values for x, which we write
symbolically as
x = ±
√
−t ±(Ax +B).
1.51. Remark. In the case of cubic and quartic polynomials over C we can obtain all the
roots by repeatedly taking square or cube roots (or radicals). Consequently such polynomials are
said to be solvable by radicals. Later we will see that this is not true in general for polynomials
of degree at least 5; this is one of the great early successes of this theory.
1.5. Automorphisms of rings and ﬁelds
1.52. Definition. Let R be a ring and R
0
⊆ R a subring.
• An automorphism of R is a ring isomorphism α: R −→ R. The set of all such auto
morphisms is denoted Aut(R).
• An automorphism of R over R
0
is a ring isomorphism α: R −→ R for which α(r) = r
whenever r ∈ R
0
. The set of all automorphisms of R over R
0
is denoted Aut
R
0
(R).
1.53. Proposition. For a ring R with a subring R
0
⊆ R, Aut(R) and Aut
R
0
(R) form
groups under composition of functions.
Proof. The composition α ◦ β of two automorphisms α, β: R −→ R is also an automor
phism of R as is the inverse of α. The identity function id = id
R
: R −→ R is an automorphism.
Hence Aut(R) forms a group under composition. The argument for Aut
R
0
(R) is similar.
1.54. Proposition. Let R be one of the core rings Z or Z/n with n > 1. Then
(i) The only automorphism of R is the identity, i.e., Aut(R) = {id}.
(ii) If S is a ring containing a core ring R and α ∈ Aut(S), then α restricts to the identity
on R, i.e., α(r) = r for all r ∈ R. Hence, Aut(S) = Aut
R
(S).
Proof. (i) For such a core ring R, every element has the form k1 for some k ∈ Z. For an
automorphism α of R,
α(k1) =
α(1) +· · · +α(1)
. .. .
k
if k > 0,
−(α(1) +· · · +α(1)
. .. .
−k
) if k < 0,
α(0) if k = 0
=
1 +· · · + 1
. .. .
k
if k > 0,
−(1 +· · · + 1
. .. .
−k
) if k < 0,
0 if k = 0
=k1.
Thus α = id.
(ii) For α ∈ Aut(S), α(1) = 1 and a similar argument to that for (i) shows that α(r) = r for all
r ∈ R.
19
1.55. Proposition. Let D be an integral domain and α: D −→ D be an automorphism.
Then the induced homomorphism gives an automorphism α
∗
: Fr(D) −→ Fr(D).
Proof. Given α, the induced homomorphism α
∗
: Fr(D) −→ Fr(D) exists and we need
to show it has an inverse. The inverse automorphism α
−1
: D −→ D also gives rise to an
induced homomorphism (α
−1
)
∗
: Fr(D) −→ Fr(D). Since α
−1
◦ α = id = α◦ α
−1
, we can apply
Corollary 1.20 to show that
(α
−1
)
∗
◦ (α)
∗
= id = (α)
∗
◦ (α
−1
)
∗
.
Hence (α)
∗
is invertible with inverse (α
−1
)
∗
.
1.56. Corollary. There is a monomorphism of groups
( )
∗
: Aut(D) −→ Aut(Fr(D)); α −→ α
∗
.
1.57. Example. The ﬁeld of fractions of the ring of integers Z is the ﬁeld of rationals Q.
The homomorphism
( )
∗
: Aut(Z) −→ Aut(Q); α −→ α
∗
is an isomorphism and hence Aut(Q) = {id}.
Combining this example with Proposition 1.54(ii) we obtain another useful result.
1.58. Proposition. Let k be one of the prime ﬁelds Q or F
p
with p > 0 prime. If R is
a ring containing k as a subring, then every automorphism of R restricts to the identity on k,
i.e., Aut(R) = Aut
k
(R).
Recalling Deﬁnition 1.36, we have an example which shows that the monomorphism of
Corollary 1.56 need not be an epimorphism. Here we take D = Q[X] and Fr(Q[X]) = Q(X).
1.59. Example. The homomorphism
( )
∗
: Aut(Q[X]) −→ Aut(Q(X)); α −→ α
∗
is a monomorphism but it is not an epimorphism since there is an automorphism
γ : Q(X) −→Q(X); γ(f(X)) = f(1/X)
which sends X ∈ Q[X] ⊆ Q(X) to 1/X / ∈ Q[X] and so does not restrict to an automorphism of
Q[X].
Let k be a ﬁeld. The group of invertible 2 × 2 matrices over k is the 2 × 2 general linear
group over k,
GL
2
(k) =
¸
a
11
a
12
a
21
a
22
¸
: a
ij
∈ k, a
11
a
22
−a
12
a
21
̸= 0
¸
The scalar matrices form a normal subgroup
Scal
2
(k) = {diag(t, t) : t ∈ k, t ̸= 0} ▹ GL
2
(k).
The quotient group is called the 2 ×2 projective general linear group over k,
PGL
2
(k) = GL
2
(k)/ Scal
2
(k).
20
Notice that GL
2
(k) has another interesting subgroup called the aﬃne subgroup,
Aﬀ
1
(k) =
¸
a b
0 1
¸
: a, b ∈ k, a ̸= 0
¸
GL
2
(k).
1.60. Example. Let k be a ﬁeld and X an indeterminate. Then Aut
k
(k[X]) and hence
Aut
k
(k(X)), contains a subgroup isomorphic to Aﬀ
1
(k). In fact, Aut
k
(k[X])
∼
= Aﬀ
1
(k).
Proof. We begin by showing that to each aﬃne matrix
A =
¸
a b
0 1
¸
∈ Aﬀ
1
(k)
there is an associated automorphism α
A
: k[X] −→k[X].
For this we use the element aX + b ∈ k[X] together with the extension result of Theo
rem 1.22(i) to obtain a homomorphism α
A
: k[X] −→ k[X] with α
A
(X) = aX + b. Using the
inverse matrix
A
−1
=
¸
a
−1
−a
−1
b
0 1
¸
we similarly obtain a homomorphism α
A
−1 : k[X] −→k[X] for which
α
A
−1(X) = a
−1
X −a
−1
b.
Using the same line of argument as in the proof of Proposition 1.55 (or doing a direct calculation)
we see that α
A
−1 is the inverse of α
A
an so α
A
∈ Aut
k
(k[X]). It is straightforward to check
that for A
1
, A
2
∈ Aﬀ
1
(k),
α
A
2
A
1
= α
A
1
◦ α
A
2
,
(note the order!) hence there is a homomorphism of groups
Aﬀ
1
(k) −→ Aut
k
(k[X]); A −→ α
A
−1,
which is easily seen to be a monomorphism. Composing with ( )
∗
we see that there is a
monomorphism Aﬀ
1
(k) −→ Aut
k
(k(X)). In fact, this is also an epimorphism and we leave the
proof of this as an exercise.
1.61. Example. Let k be a ﬁeld and X an indeterminate. Then
(i) Aut
k
(k(X)) contains a subgroup isomorphic to PGL
2
(k).
(ii) In fact, Aut
k
(k(X))
∼
= PGL
2
(k).
Proof. (i) We begin by showing that to each invertible matrix
A =
¸
a
11
a
12
a
21
a
22
¸
∈ GL
2
(k)
there is an associated automorphism α
A
: k(X) −→k(X).
We begin by choosing the element (a
11
X +a
12
)/(a
21
X +a
22
) ∈ k(X) and then using Theo
rem 1.22(i) to obtain a homomorphismk[X] −→k(X) that sends X to (a
11
X+a
12
)/(a
21
X+a
22
).
By applying ( )
∗
to this we obtain a homomorphism (known as a fractional linear transforma
tion) α
A
: k(X) −→k(X) for which
α
A
(X) =
a
11
X +a
12
a
21
X +a
22
.
21
Again we ﬁnd that
α
A
2
A
1
= α
A
1
◦ α
A
2
.
There is an associated homomorphism of groups GL
2
(k) −→ Aut
k
(k(X)) sending A to α
A
−1
.
However, this is not an injection in general since for each scalar matrix diag(t, t),
α
diag(t,t)
(X) =
tX
t
= X,
showing that α
diag(t,t)
is the identity function.
In fact it is easy to see that Scal
2
(k)▹GL
2
(k) is the kernel of this homomorphism. Therefore
passing to the quotient PGL
2
(k) = GL
2
(k)/ Scal
2
(k) we obtain a monomorphism PGL
2
(k) −→
Aut
k
(k(X)). There is one case where Scal
2
(k) is the trivial group, namely k = F
2
.
(ii) To show that every automorphism of k(X) is a fractional linear transformation is less
elementary. We give a sketch proof for the case of k = C; actually this argument can be modiﬁed
to work for any algebraically closed ﬁeld, but an easy argument then shows the general case.
Let α ∈ Aut
C
(C(X)). There is an associated rational (hence meromorphic) function f given
by z −→ f(z), where α(X) = f(X), deﬁned on C with the poles of f deleted. If we write
f(X) =
p(X)
q(X)
where p(X), q(X) ∈ C[X] have no common factors of positive degree, then the order of f(X) is
ord f = max{deg p(X), deg q(X)}.
Now let c ∈ C. Then the number of solutions counted with algebraic multiplicity of the equation
f(z) = c turns out to be ord f. Also, if deg p(X) deg q(X) then the number of poles of f
counted with algebraic multiplicity is also ord f. Finally, if deg p(X) > deg q(X) then we can
write
f(X) = p
1
(X) +
p
0
(X)
q(X)
,
where p
0
(X), p
1
(X) ∈ C[X] and deg p
0
(X) < deg q(X). Then the number of poles of f counted
with algebraic multiplicity is
deg p
1
(X) + ord
p
0
q
.
Now it is easy to see that since α is invertible so is the function f. But this can only happen
if the function f is injective which means that all of these numbers must be 1, hence ord f = 1.
Thus
f(X) =
aX +b
cX +d
̸= constant
and the matrix
¸
a b
c d
¸
must be invertible.
Clearly not every fractional linear transformation α
A
: k(X) −→k(X) maps polynomials to
polynomials so ( )
∗
: Aut
k
(k[X]) −→ Aut
k
(k(X)) is not an epimorphism.
Now we turn to a more familiar ﬁeld R, the real numbers.
1.62. Proposition. The only automorphism of the ﬁeld R is the identity function, hence
Aut(R) = {id}.
22
Proof. First we note that Q ⊆ R is a subring and if α ∈ Aut(R) then α(q) = q for q ∈ Q
by Example 1.57.
We recall from Analysis that the rational numbers are dense in the real numbers in the
sense that each r ∈ R can be expressed as a limit r = lim
n→∞
q
n
, where q
n
∈ Q. Then for a
continuous function f : R −→R, its value at r depends on its values on Q since
f(r) = f( lim
n→∞
q
n
) = lim
n→∞
f(q
n
).
We will show that an automorphism α ∈ Aut(R) is continuous.
First recall that for x, y ∈ R,
x < y ⇐⇒ 0 < y −x ⇐⇒ y −x = t
2
for some nonzero t ∈ R.
Now for α ∈ Aut(R) and s ∈ R, we have α(s
2
) = α(s)
2
. Hence,
x < y =⇒ α(y) −α(x) = α(t)
2
for some nonzero t ∈ R =⇒ α(x) < α(y).
So α preserves order and ﬁxes rational numbers.
Now let x ∈ R and ε > 0. Then we can choose a rational number q such that 0 < q ε.
Taking δ = q we ﬁnd that for y ∈ R with y −x < δ (i.e., −δ < y −x < δ) we have
−δ = α(−δ) < α(y) −α(x) < α(δ) = δ,
hence
α(y) −α(x) < δ ε.
This shows that α is continuous at x.
Thus every automorphism of R is continuous function which ﬁxes all the rational numbers,
hence it must be the identity function.
1.63. Remark. If we try to determine Aut(C) the answer turns out to be much more
complicated. It is easy to see that complex conjugation ( ): C −→C is an automorphism of C
and ﬁxes every real number, i.e., ( ) ∈ Aut
R
(C); in fact, Aut
R
(C) = {id, ( )}. However, it is
not true that every α ∈ Aut(C) ﬁxes every real number! The automorphism group Aut(C) is
actually enormous but it is hard to ﬁnd an explicit element other than id and ( ). Note that
given an automorphism α ∈ Aut(C), the composition α ◦ ( ) ◦ α
−1
is also self inverse, so there
are many elements of order 2 in the group Aut(C).
Exercises on Chapter 1
1.1. Let R be a ring. Show that
{n ∈ Z : n > 0 and n1 = 0} = {n ∈ Z : n > 0 and nr = 0 for all r ∈ R}.
Deduce that if char R > 0 then these sets are nonempty and
char R = min{n ∈ Z : n > 0 and nr = 0 for all r ∈ R}.
1.2. Let R be an integral domain.
(a) Show that every subring S ⊆ R is also an integral domain. What is the relationship
between char S and char R?
23
(b) If R is a ﬁeld, give an example to show that a subring of R need not be a ﬁeld.
1.3. For each of the following rings R, ﬁnd the characteristic char R and the characteristic
subring of R. Determine which of these rings is an integral domain. In (b) and (c), A is an
arbitrary commutative ring.
(a) Any subring R ⊆ C.
(b) The polynomial ring R = A[X].
(c) The ring of n ×n matrices over A,
R = Mat
n
(A) =
a
11
. . . a
1n
.
.
.
.
.
.
.
.
.
a
n1
. . . a
nn
¸
¸
¸
¸
: a
ij
∈ A
.
1.4. If R is a commutative ring with unit containing the prime ﬁeld F
p
for some prime p > 0,
show that the function φ: R −→ R given by φ(t) = t
p
, deﬁnes a ring homomorphism. Give
examples to show that φ need not be surjective or injective.
1.5. Let R and S be rings with unity and Q▹ S a prime ideal.
(a) If φ: R −→ S is a ring homomorphism, show that
φ
−1
Q = {r ∈ R : φ(r) ∈ Q} ⊆ R
is a prime ideal of R.
(b) If R ⊆ S is a subring, show that Q∩ R is a prime ideal of R.
(c) If the word ‘prime’ is replaced by ‘maximal’ throughout, are the results in parts (a)
and (b) still true? [Hint: look for a counterexample.]
(d) If R ⊆ S is a subring and P ▹ R is a maximal ideal, suppose that Q▹ S is a prime ideal
for which P ⊆ Q. Show that Q∩ R = P.
1.6. Let k be a ﬁeld, R be a ring with unit and let φ: k −→ R be a ring homomorphism. Show
that φ is a monomorphism.
1.7. Consider the sets
Z(i) = {u +vi : u, v ∈ Z} ⊆ C, Q(i) = {u +vi : u, v ∈ Q} ⊆ C.
(a) Show that Z(i) and Q(i) are subrings of C. Also show that Z(i) is an integral domain,
Q(i) is a ﬁeld and Z(i) is a subring of Q(i).
(b) Show that the inclusion homomorphism inc: Z(i) −→Q(i) extends to a monomorphism
inc
∗
: Fr(Z(i)) −→Q(i).
(c) Show that inc
∗
is an isomorphism, so Fr(Z(i)) = Q(i).
1.8. Let R be a commutative ring.
(a) If a, b ∈ R, show that there is a unique ring homomorphism ψ
a,b
: R[X] −→ R[X] for
which ψ
a,b
(r) = r if r ∈ R and ψ
a,b
(X) = aX +b. If c, d ∈ R, determine ψ
a,b
◦ ψ
c,d
. If
a is a unit, show that ψ
a,b
is an isomorphism and ﬁnd its inverse.
(b) Now suppose that R = k is a ﬁeld and a, b ∈ k with a ̸= 0. Prove the following.
(i) If f(X) ∈ k[X], the deg ψ
a,b
(f(X)) = deg f(X).
(ii) If p(X) ∈ k[X] is a prime then so is ψ
a,b
(p(X)).
24
(iii) If p(X) ∈ k[X] is an irreducible then so is ψ
a,b
(p(X)).
1.9. Let k be a ﬁeld and k[[X]] be the set consisting of all power series
∞
∑
k=0
a
k
X
k
= a
0
+a
1
X +· · · +a
k
X
k
+· · · ,
with a
k
∈ k.
(a) Show that this can be made into an integral domain containing k[X] as a subring by
deﬁning addition and multiplication in the obvious way.
(b) Show that
∑
∞
k=0
a
k
X
k
∈ k[[X]] is a unit if and only if a
0
̸= 0.
(c) Show that Fr(k[[X]]) consists of all ﬁnitetailed Laurent series
∞
∑
k=ℓ
a
k
X
k
= a
ℓ
X
ℓ
+a
ℓ+1
X
ℓ+1
+· · · +a
k
X
k
+· · ·
for some ℓ ∈ Z and a
k
∈ k.
1.10. Taking k = Q, ﬁnd the quotient and remainder when performing long division of f(X) =
6X
4
−6X
3
+ 3X
2
−3X −2 by d(X) = 2X
3
+X + 3.
1.11. Taking k = F
3
, ﬁnd the quotient and remainder when performing long division of
f(X) = 2X
3
+ 2X
2
+X + 1 by d(X) = 2X
3
+ 2X.
1.12. Let p > 0 be a prime. Suppose that f(X) = a
0
+ a
1
X + · · · + a
n
X
n
∈ Z[X] with
p a
n
and that f(X) ∈ F
p
[X] denotes the polynomial obtained by reducing the coeﬃcients of
f(X) modulo p. If f(X) is irreducible, show that f(X) is irreducible. Which of the following
polynomials in Z[X] is irreducible?
X
3
−X + 1, X
3
+ 2X + 1, X
3
+X −1, X
5
−X + 1, X
5
+X −1, 5X
3
−10X +X
2
−2.
1.13. Find generators for each of the following ideals:
I
1
= {f(X) ∈ Q[X] : f(i) = 0} ▹ Q[X], I
2
= {f(X) ∈ Q[X] : f(
√
2 i) = 0} ▹ Q[X],
I
3
= {f(X) ∈ Q[X] : f(
√
2) = 0} ▹ Q[X], I
4
= {f(X) ∈ R[X] : f(
√
2) = 0} ▹ R[X],
I
5
= {f(X) ∈ R[X] : f(
√
2 i) = 0} ▹ R[X], I
6
= {f(X) ∈ R[X] : f(ζ
3
) = 0} ▹ R[X].
1.14. Consider the inclusion inc: Q −→C and its extension to ε
√
2
: Q[X] −→C.
Determine the image ε
√
2
Q[X] ⊆ C. What is ε
−
√
2
Q[X] ⊆ C? Find ker ε
√
2
▹Q[X] and
ker ε
−
√
2
▹Q[X]; are these maximal ideals?
1.15. Let ω = (−1 +
√
3i)/2 ∈ C. Consider the inclusion inc: Q −→ C and its extension
to ε
ω
: Q[X] −→ C. Determine the image ε
ω
Q[X] ⊆ C. Determine ker ε
ω
▹Q[X] and decide
whether it is maximal. Find another evaluation homomorphism with the same kernel and image.
1.16. Consider the inclusion inc: Q −→ C and its extension to ε
α
: Q[X] −→ C where α is
one of the 4 complex roots of the polynomial f(X) = X
4
− 2 ∈ Q[X]. Determine the image
ε
α
Q[X] ⊆ C and the ideal ker ε
α
▹Q[X]; is the latter ideal maximal? What happens if α is
replaced by one of the other roots of f(X)?
Repeat this problem starting with the inclusion of the real numbers into the complex num
bers inc: R −→C and ε
α
: R[X] −→C.
25
1.17. Use Cardan’s method to ﬁnd the complex roots of the polynomial
f(X) = X
3
−9X
2
+ 21X −5.
1.18. Consider the real numbers
α =
3
√
10 +
√
108 +
3
√
10 −
√
108, β =
3
√
1 +
2
3
√
7
3
+
3
√
1 −
2
3
√
7
3
.
Find rational cubic polynomials f(X) and g(X) for which f(α) = 0 = g(β). Hence determine
these real numbers.
1.19. Prove the ﬁnal part of Example 1.60 by showing that there is an isomorphism of groups
Aﬀ
1
(k)
∼
= Aut
k
(k[X]).
1.20. Let k be any ﬁeld. Consider the 6 automorphisms α
j
: k(X) −→ k(X) (j = 1, . . . , 6)
deﬁned by
α
1
(f(X)) = f(X), α
2
(f(X)) = f(1 −X), α
3
(f(X)) = f(1/X),
α
4
(f(X)) = f((X −1)/X), α
5
(f(X)) = f(1/(1 −X)), α
6
(f(X)) = f(X/(X −1)).
Show that the set consisting of these elements is a subgroup Γ
k
Aut
k
(k(X)) isomorphic to
the symmetric group S
3
. When k = F
2
, show that Γ
k
∼
= GL
2
(k).
1.21. Determine the cyclotomic polynomial Φ
20
(X).
1.22. Let p > 0 be a prime.
(a) Show that for k 1, the cyclotomic polynomial Φ
p
k(X) satisﬁes
Φ
p
k(X) = Φ
p
(X
p
k−1
)
and has as its complex roots the primitive p
k
th roots of 1.
(b) Show that Φ
p
k(X) ∈ Q[X] is irreducible.
(c) Generalize part (a) to show that if n = p
r
1
1
· · · p
r
k
k
is the prime power factorization of n
with the p
i
being distinct primes and r
i
> 0, then
Φ
n
(X) = Φ
p
1
···p
k
(X
p
r
1
−1
1
···p
r
k
−1
k
).
1.23. For n 2, show that
X
φ(n)
Φ
n
(X
−1
) = Φ
n
(X).
1.24. Show that for n 1, ζ
n
+ζ
−1
n
= 2 cos(2π/n).
Find expressions for ζ
5
+ζ
−1
5
and ζ
2
5
+ζ
−2
5
in terms of cos(2π/5). Hence ﬁnd a rational polynomial
which has cos(2π/5) as a root.
1.25. Let p > 0 be a prime and K be a ﬁeld with char K = p.
(a) Show that if ζ ∈ K is a pth root of 1 then ζ = 1. Deduce that if m, n > 0 and p n,
then every np
m
th root of 1 in K is an nth root of 1.
(b) If a ∈ K, show that the polynomial X
p
−a ∈ K[X] has either no roots or exactly one
root in K.
26
CHAPTER 2
Fields and their extensions
2.1. Fields and subﬁelds
2.1. Definition. Let K and L be ﬁelds and suppose that K ⊆ L is a subring. Then we say
that K is a subﬁeld of L; L is also said to be an extension (ﬁeld) of K. We write K L or
L/K to indicate this, and write K < L if K is a proper subﬁeld of L, i.e., if K ̸= L.
An important fact about an extension of ﬁelds L/K is that L is a Kvector space whose
addition is the addition in the ﬁeld L while scalar multiplication is deﬁned by
u · x = ux (u ∈ K, x ∈ L).
2.2. Definition. We will call dim
K
L the degree or index of the extension L/K and use the
notation [L : K] = dim
K
L. An extension of ﬁelds L/K is ﬁnite (dimensional ) if [L : K] < ∞,
otherwise it is inﬁnite (dimensional ).
2.3. Example. Show that the extension C/R is ﬁnite, while R/Q and C/Q are both inﬁnite.
Solution. We have
C = {x +yi : x, y ∈ R},
so 1, i span C as a vector space over R. Since i / ∈ R, these elements are also linearly independent
over R and therefore they form a basis, whence [C : R] = 2. The inﬁniteness of R/Q and C/Q are
consequences of the fact that any ﬁnite dimensional vector space over Q is countable, however
R and C are uncountable. A basis for the Qvector space R is known as a Hamel basis.
2.4. Example. Consider the extension Q(
√
2)/Q where
Q(
√
2) = {x +y
√
2 : x, y ∈ Q}.
Show that [Q(
√
2) : Q] = 2.
Solution. The elements 1,
√
2 clearly span the Qvector space Q(
√
2). Now recall that
√
2 / ∈ Q. If the elements 1,
√
2 were linearly dependent we would have u + v
√
2 = 0 for some
u, v ∈ Q not both zero; in fact it is easy to see that we would then also have u, v both nonzero.
Thus we would have
√
2 = −
u
v
∈ Q,
which we know to be false. Hence 1,
√
2 are linearly independent and so form a basis for Q(
√
2)
over Q and [Q(
√
2) : Q] = 2.
If we have two extensions L/K and M/L then it is a straightforward to verify that K M
and so we have another extension M/K.
2.5. Definition. Given two extensions L/K and M/L, we say that L/K is a subextension
of M/K and sometimes write L/K M/K.
27
2.6. Theorem. Let L/K be a subextension of M/K.
(i) If one or both of the dimensions [L : K] or [M : L] is inﬁnite then so is [M : K].
(ii) If the dimensions [L : K] and [M : L] are both ﬁnite then so is [M : K] and
[M : K] = [M : L] [L : K].
Proof. (i) If [M : K] is ﬁnite, choose a basis m
1
, . . . , m
r
of M over K. Now any element
u ∈ M can be expressed as
u = t
1
m
1
+· · · +t
r
m
r
,
where t
1
, . . . , t
r
∈ K; but since K ⊆ L, this means that m
1
, . . . , m
r
spans M over L and so
[M : L] < ∞. Also L is a Kvector subspace of the ﬁnite dimensional Kvector space M, hence
[L : K] < ∞.
(ii) Setting r = [L : K] and s = [M : L], choose a basis ℓ
1
, . . . , ℓ
r
of L over K and a basis
m
1
, . . . , m
s
of M over L.
Now let v ∈ M. Then there are elements y
1
, . . . , y
s
∈ L for which
v = y
1
m
1
+· · · +y
s
m
s
.
But each y
j
can be expressed in the form
y
j
= x
1j
ℓ
1
+· · · +x
rj
ℓ
r
for suitable x
ij
∈ K. Hence,
v =
s
∑
j=1
r
∑
i=1
x
ij
ℓ
i
m
j
=
s
∑
j=1
r
∑
i=1
x
ij
(ℓ
i
m
j
),
where each coeﬃcient x
ij
is in K. Thus the elements ℓ
i
m
j
(i = 1, . . . , r, j = 1, . . . , s) span the
Kvector space M.
Now suppose that for some t
ij
∈ K we have
s
∑
j=1
r
∑
i=1
t
ij
(ℓ
i
m
j
) = 0.
On collecting terms we obtain
s
∑
j=1
r
∑
i=1
t
ij
ℓ
i
m
j
= 0,
where each coeﬃcient
∑
r
i=1
t
ij
ℓ
i
is in L. By the linear independence of the m
j
over L, this
means that for each j,
r
∑
i=1
t
ij
ℓ
i
= 0.
By the linear independence of the ℓ
i
over K, each t
ij
= 0.
Hence the ℓ
i
m
j
form a basis of M over K and so
[M : K] = rs = [M : L] [L : K].
28
We will often indicate subextensions in diagrammatic form where larger ﬁelds always go
above smaller ones and the information on the lines indicates dimensions
M
[M:L]
[M:K]=[M:L] [L:K]
L
[L:K]
K
We often suppress ‘composite’ lines such as the dashed one. Such towers of extensions are our
main objects of study. We can build up sequences of extensions and form towers of arbitrary
length. Thus, if L
1
/K, L
2
/L
1
, . . . , L
k
/L
k−1
is a such a sequence of extensions, there is a
diagram
L
k
L
k−1
L
1
K
2.2. Simple and ﬁnitely generated extensions
2.7. Definition. Let F be a ﬁeld and K F. Given elements u
1
, . . . , u
r
∈ F we set
K(u
1
, . . . , u
r
) =
∩
KLF
u
1
,...,u
r
∈L
L
which is the smallest subﬁeld in F that contains K and the elements u
1
, . . . , u
r
. The ex
tension K(u
1
, . . . , u
r
)/K is said to be generated by the elements u
1
, . . . , u
r
; we also say that
K(u
1
, . . . , u
r
)/K is a ﬁnitely generated extension of K. An extension of the form K(u)/K is
called a simple extension of K with generator u.
We can extend this to the case of an inﬁnite sequence u
1
, . . . , u
r
, . . . in F and denote by
K(u
1
, . . . , u
r
, . . .) F the smallest extension ﬁeld of K containing all the elements u
r
.
It can be shown that
(2.1) K(u
1
, . . . , u
r
) =
f(u
1
, . . . , u
r
)
g(u
1
, . . . , u
r
)
∈ F : f(X
1
, . . . , X
r
), g(X
1
, . . . , X
r
) ∈ K[X
1
, . . . , X
r
], g(u
1
, . . . , u
r
) ̸= 0
.
Reordering the u
i
does not change K(u
1
, . . . , u
n
).
29
2.8. Proposition. Let K(u)/K and K(u, v)/K(u) be simple extensions. Then
K(u, v) = K(u)(v) = K(v)(u).
More generally,
K(u
1
, . . . , u
n
) = K(u
1
, . . . , u
n−1
)(u
n
)
and this is independent of the order of the sequence u
1
, . . . , u
n
.
2.9. Theorem. For a simple extension K(u)/K, exactly one of the following conditions
holds.
(i) The evaluation at u homomorphism ε
u
: K[X] −→ K(u) is a monomorphism and on
passing to the fraction ﬁeld gives an isomorphism (ε
u
)
∗
: K(X) −→ K(u). In this case,
K(u)/K is inﬁnite and u is said to be transcendental over K.
(ii) The evaluation at u homomorphism ε
u
: K[X] −→ K(u) has a nontrivial kernel
ker ε
u
= (p(X)) where p(X) ∈ K[X] is an irreducible monic polynomial of positive de
gree and the quotient homomorphism ¯ ε
u
: K[X]/(p(X)) −→ K(u) is an isomorphism.
In this case K(u)/K is ﬁnite with [K(u) : K] = deg p(X) and u is said to be algebraic
over K.
Proof. (i) If ker ε
u
= (0), all that needs checking is that (ε
u
)
∗
is an epimorphism; but as
u is in the image of (ε
u
)
∗
this is obvious.
(ii) When ker ε
u
̸= (0), Theorem 1.31(iv) implies that the image of ε
u
is a subﬁeld of K(u) and
since it contains u it must equal K(u). Hence ¯ ε
u
is an isomorphism. Using Long Division, we
ﬁnd that every element of K[X]/(p(X)) can be uniquely expressed as a coset of the form
f(X) + (p(X)),
where deg f(X) < deg p(X). Hence every element of K[X]/(p(X)) can be uniquely expressed
as a linear combination over K of the d cosets
1 + (p(X)), X + (p(X)), X
2
+ (p(X)), . . . , X
d−1
+ (p(X)),
where d = deg p(X). Via the isomorphism ¯ ε
u
under which ¯ ε
u
(X
k
+ (p(X))) = u
k
, we see that
the elements 1, u, . . . , u
d−1
form a basis for K(u) over K.
2.10. Example. For the extension Q(
√
2,
√
3)/Q we have [Q(
√
2,
√
3) : Q] = 4.
Proof. By Example 2.4 we know that [Q(
√
2) : Q] = 2. We have the following tower of
extensions.
Q(
√
2,
√
3)
[Q(
√
2,
√
3):Q(
√
2)]
[Q(
√
2,
√
3):Q]=2[Q(
√
2,
√
3):Q(
√
2)] Q(
√
2)
2
Q
We will show that [Q(
√
2,
√
3) : Q(
√
2)] = 2.
30
Notice that if u ∈ Q(
√
2,
√
3) = Q(
√
2)(
√
3) then u = a + b
√
3 for some a, b ∈ Q(
√
2),
so 1,
√
3 span Q(
√
2,
√
3) over Q(
√
2). But if these are linearly dependent then
√
3 ∈ Q(
√
2).
Writing
√
3 = v +w
√
2
with v, w ∈ Q, we ﬁnd that
v
2
+ 2w
2
+ 2vw
√
2 = 3 ∈ Q,
and hence 2vw
√
2 ∈ Q. The possibilities v = 0 or w = 0 are easily ruled out, while v, w ̸= 0
would implies that
√
2 ∈ Q which is false. So 1,
√
3 are linearly independent over Q(
√
2)
and therefore form a basis of Q(
√
2,
√
3). This shows that [Q(
√
2,
√
3) : Q(
√
2)] = 2 and so
[Q(
√
2,
√
3) : Q] = 4.
2.11. Remark. There are some other subﬁelds of Q(
√
2,
√
3) which are conveniently dis
played in the following diagram.
Q(
√
2,
√
3)
2
.
.
.
.
.
.
.
.
.
.
2
2
Q(
√
2)
2
Q(
√
3)
2
Q(
√
6)
2
.
.
.
.
.
.
.
.
.
.
.
.
Q
One idea in the veriﬁcation of Example 2.10 can be extended to provide a useful general
result whose proof is left as an exercise.
2.12. Proposition. Let p
1
, . . . , p
n
be a sequence of distinct primes p
i
> 0. Then
√
p
n
/ ∈ Q(
√
p
1
, . . . ,
√
p
n−1
).
Hence [Q(
√
p
1
, . . . ,
√
p
n
) : Q(
√
p
1
, . . . ,
√
p
n−1
)] = 2 and [Q(
√
p
1
, . . . ,
√
p
n
) : Q] = 2
n
.
2.13. Example. For the extension Q(
√
2, i)/Q we have [Q(
√
2, i) : Q] = 4.
Proof. We know that [Q(
√
2) : Q] = 2. Also, i / ∈ Q(
√
2) since i is not real and Q(
√
2) R.
Since i
2
+ 1 = 0, we have Q(
√
2, i) = Q(
√
2)(i) and [Q(
√
2, i) : Q(
√
2)] = 2. Using the formula
[Q(
√
2, i) : Q] = [Q(
√
2, i) : Q(
√
2)] [Q(
√
2) : Q],
we obtain [Q(
√
2, i) : Q] = 4.
31
This example also has several other subﬁelds, with only Q(
√
2) = Q(
√
2, i) ∩ R being a
subﬁeld of R.
C
2
∞ R
∞ Q(
√
2, i)
2
2
2
Q(
√
2)
2
Q(i)
2
Q(
√
2 i)
2
.
.
.
.
.
.
.
.
.
.
.
Q
2.14. Example. For n 1, let E
n
= Q(2
1/n
) R, where 2
1/n
∈ R denotes the positive real
nth root of 2.
(i) Show that [E
n
: Q] = n.
(ii) If m 1 with m  n, show that E
m
E
n
and determine [E
n
: E
m
].
(iii) If m, n are coprime, show that E
mn
= Q(2
1/m
, 2
1/n
).
Solution. (i) Consider the evaluation homomorphism ε
2
1/n : Q[X] −→ E
n
. Applying the
Eisenstein Test 1.38 using the prime 2 to the polynomial X
n
−2 ∈ Z[X], we ﬁnd that
ker ε
2
1/n = (X
n
−2) ▹ Q[X],
and the induced homomorphism ¯ ε
2
1/n : Q[X]/(X
n
− 2) −→ E
n
is an isomorphism. Hence
[E
n
: Q] = n.
(ii) Since n/m is an integer,
2
1/m
= (2
1/n
)
n/m
∈ E
n
,
so
E
m
= Q(2
1/m
) ⊆ E
n
.
By Theorem 2.6 we have
n = [E
n
: Q] = [E
n
: E
m
] [E
m
: Q] = m[E
n
: E
m
],
whence [E
n
: E
m
] = n/m.
(iii) By (ii) we have E
m
E
mn
and E
n
E
mn
, hence Q(2
1/m
, 2
1/n
) E
mn
. As gcd(m, n) = 1,
there are integers r, s for which rm+sn = 1 and so
1
mn
=
rm+sn
mn
=
r
n
+
s
m
.
This shows that
2
1/mn
= (2
1/n
)
r
(2
1/m
)
s
∈ Q(2
1/m
, 2
1/n
),
whence E
mn
Q(2
1/m
, 2
1/n
). Combining these inclusions we obtain E
mn
= Q(2
1/m
, 2
1/n
).
32
Exercises on Chapter 2
2.1. Let p ∈ N be an prime. Show that the extension Q(
√
p)/Q has [Q(
√
p) : Q] = 2.
2.2. Let p, q > 0 be distinct primes. Show that [Q(
√
p,
√
q) : Q(
√
p)] = 2.
2.3. Prove Proposition 2.12 by induction on n.
2.4. Let K a ﬁeld with char K ̸= 2 and suppose that L/K is an extension. If a, b ∈ K are
distinct, suppose that u, v ∈ L satisfy u
2
= a and v
2
= b. Show that K(u, v) = K(u +v).
[Hint: ﬁrst show that u±v ̸= 0 and deduce that u−v ∈ K(u+v); then show that u, v ∈ K(u+v).]
2.5. Show that [Q(i) : Q] = 2.
2.6. Show that [Q(
√
3, i) : Q] = 4. Find the three subﬁelds L Q(
√
3, i) with [L : Q] = 2 and
display their relationship in a diagram, indicating which ones are subﬁelds of R.
2.7. Let ζ
5
= e
2πi/5
∈ C.
(a) Explain why [Q(ζ
5
) : Q] = 4.
(b) Show that cos(2π/5), sin(2π/5) i ∈ Q(ζ
5
).
(c) Show that for t ∈ R,
cos 5t = 16 cos
5
t −20 cos
3
t + 5 cos t.
(d) Show that the numbers cos(2kπ/5) with k = 0, 1, 2, 3, 4 are roots of the polynomial
f(X) = 16X
5
−20X
3
+ 5X −1 = (X −1)(4X
2
+ 2X −1)
2
and deduce that [Q(cos(2π/5)) : Q] = 2.
(e) Display the relationship between the ﬁelds Q, Q(cos(2π/5)), and Q(ζ
5
) in a suitable
diagram.
2.8. This question is for those who like lots of calculation or using Maple. Let ζ
7
= e
2πi/7
∈ C.
(a) Explain why [Q(ζ
7
) : Q] = 6.
(b) Show that cos(2π/7), sin(2π/7) i ∈ Q(ζ
7
).
(c) Show
cos 7t = 64 cos
7
t −112 cos
5
t + 56 cos
3
t −7 cos t.
Show that the numbers cos(2kπ/7) with k = 0, 1, . . . , 6 are roots of the polynomial
f(X) = 64X
7
−112X
5
+ 56X
3
−7X −1 = (X −1)(8X
3
+ 4X
2
−4X −1)
2
and deduce that [Q(cos(2π/7)) : Q] = 3.
(d) Show that sin(2π/7) i is a root of
g(X) = 64X
7
+ 112X
5
+ 56X
3
+ 7X = X(64X
6
+ 112X
4
+ 56X
2
+ 7)
and that 64X
6
+112X
4
+56X
2
+7 ∈ Q[X] is irreducible. What is [Q(sin(2π/7) i) : Q]?
(e) Display the relationship between the ﬁelds Q, Q(cos(2π/7)), Q(sin(2π/7) i) and Q(ζ
7
)
in a diagram.
(f) Is i ∈ Q(ζ
7
)?
2.9. In this question we continue to consider the situation described in Example 2.14.
33
(a) Show that
Aut
Q
(E
n
) =
{id} if n is odd,
{id, τ
n
}
∼
= Z/2 if n is even,
where τ
n
has composition order 2.
(b) Let E =
∪
n1
E
n
R. Show that Aut
Q
(E) = {id}.
(c) Display the 6 subﬁelds of E
12
in a diagram.
(d) Which of the subﬁelds in part (c) contain the element 2
1/2
+ 2
1/3
?
34
CHAPTER 3
Algebraic extensions of ﬁelds
3.1. Algebraic extensions
Let L/K be an extension of ﬁelds. From Theorem 2.9(ii), recall the following notion.
3.1. Definition. An element t ∈ L is algebraic over K if there is a nonzero polynomial
p(X) ∈ K[X] for which p(t) = 0.
Notice in particular that for an element t ∈ K, the polynomial p(X) = X − t ∈ K[X]
satisﬁes p(t) = 0, so t is algebraic over K.
Theorem 2.9 allows us to characterize algebraic elements in other ways.
3.2. Proposition. Let t ∈ L. Then the following conditions are equivalent.
(i) t is algebraic over K.
(ii) The evaluation homomorphism ε
t
: K[X] −→ L has nontrivial kernel.
(iii) The extension K(t)/K is ﬁnite dimensional.
3.3. Definition. If t ∈ L is algebraic over K then by Proposition 3.2,
ker ε
t
= (minpoly
K,t
(X)) ̸= (0),
where minpoly
K,t
(X) ∈ K[X] is an irreducible monic polynomial called the minimal polynomial
of t over K. The degree of minpoly
K,t
(X) is called the degree of t over K and is denoted
deg
K
t.
3.4. Proposition. If t ∈ L is algebraic over K then
[K(t) : K] = deg minpoly
K,t
(X) = deg
K
t.
Proof. This follows from Theorem 2.9(ii).
3.5. Remark. Suppose that t ∈ L is algebraic over K and that p(X) ∈ ker ε
t
with
deg p(X) = deg minpoly
K,t
(X). Then minpoly
K,t
(X)  p(X) and so
p(X) = uminpoly
K,t
(X)
for some u ∈ K. In particular, when p(X) is monic,
p(X) = minpoly
K,t
(X).
We will often use this without further comment.
3.6. Example. Consider C/Q. The minimal polynomial of
√
2 ∈ C over Q is
minpoly
Q,
√
2
(X) = X
2
−2.
35
Proof. Clearly X
2
−2 ∈ ker ε
√
2
since (
√
2)
2
−2 = 0. By Example 2.4,
deg minpoly
Q,
√
2
(X) = [Q(
√
2) : Q] = 2,
hence
minpoly
Q,
√
2
(X) = X
2
−2.
3.7. Example. Consider C/Q. The minimal polynomial of i ∈ C over Q is X
2
+ 1.
Proof. Clearly X
2
+ 1 ∈ ker ε
i
since i
2
+ 1 = 0. As [Q(i) : Q] = 2, we have
minpoly
Q,i
(X) = X
2
+ 1.
3.8. Example. Consider C/Q. Find the minimal polynomial of the primitive 6th root of
unity, ζ
6
∈ C over Q.
Solution. Recall from Example 1.44 that ζ
6
is a root of the irreducible cyclotomic poly
nomial
Φ
6
(X) = X
2
−X + 1.
Then Φ
6
(X) ∈ ker ε
ζ
6
so minpoly
Q,ζ
6
(X)  Φ
6
(X). Since Φ
6
(X) is irreducible and monic, we
must have
minpoly
Q,ζ
6
(X) = Φ
6
(X)
and so deg
Q
ζ
6
= 2.
3.9. Example. Consider C/Q. Find the minimal polynomial of
√
2 +
√
3 over Q.
Solution. Notice that
√
3 −
√
2 =
(
√
3 −
√
2)(
√
3 +
√
2)
(
√
3 +
√
2)
=
1
√
2 +
√
3
∈ Q(
√
2 +
√
3).
So we have
√
2 =
1
2
(
√
2 +
√
3) −(
√
3 −
√
2)
∈ Q(
√
2 +
√
3),
√
3 =
1
2
(
√
2 +
√
3) + (
√
3 −
√
2)
∈ Q(
√
2 +
√
3),
hence Q(
√
2,
√
3) Q(
√
2 +
√
3). Since Q(
√
2 +
√
3) Q(
√
2,
√
3) we must have
Q(
√
2 +
√
3) = Q(
√
2,
√
3).
Referring to Example 2.10 we see that
deg
Q
(
√
2 +
√
3) = 4.
Let us ﬁnd a nonzero polynomial in ker ε
√
2+
√
3
▹Q[X].
Referring to Example 2.10 or Proposition 2.12 we see that
√
2 +
√
3 / ∈ Q(
√
2), hence
deg
Q(
√
2)
(
√
2 +
√
3) = 2.
One polynomial in ker ε
√
2+
√
3
▹Q(
√
2)[X] is
(X −(
√
2 +
√
3))(X −(
√
2 −
√
3)) = X
2
−2
√
2X −1.
Since this is monic and of degree 2,
minpoly
Q(
√
2),
√
2+
√
3
(X) = X
2
−2
√
2X −1.
36
Similarly,
minpoly
Q(
√
2),−
√
2+
√
3
(X) = X
2
+ 2
√
2X −1.
Consider
p(X) = minpoly
Q(
√
2),
√
2+
√
3
(X) minpoly
Q(
√
2),−
√
2+
√
3
(X)
= (X
2
−2
√
2X −1)(X
2
+ 2
√
2X −1)
= X
4
−10X
2
+ 1.
Then p(
√
2 +
√
3) = 0 so p(X) ∈ ker ε
t
. Since deg p(X) = 4 and p(X) is monic, we have
minpoly
Q,
√
2+
√
3
(X) = X
4
−10X
2
+ 1.
3.10. Definition. Let L/K be a ﬁnite extension. An element u ∈ L for which L = K(u) is
called a primitive element for the extension L/K.
Later we will see that when char K = 0 every ﬁnite extension L/K has a primitive element.
3.11. Lemma. Let L/K be a ﬁnite extension and u ∈ L. Then u is a primitive element for
L/K if and only if deg
K
u = [L : K].
Proof. K(u) ⊆ L is a ﬁnite dimensional Kvector subspace. Then K(u) = L if and only
dim
K
K(u) = dim
K
L. Since deg
K
u = dim
K
K(u) and [L : K] = dim
K
L the result follows.
Sometimes the minimal polynomial of an element in an extension is introduced in a diﬀerent
but equivalent way.
3.12. Proposition. Let t ∈ L be algebraic over K. Then
I(t) = {f(X) ∈ K[X] : f(t) = 0} ⊆ K[X]
is an ideal which is principal and has an irreducible monic generator q(X) ∈ K[X]. In fact,
q(X) = minpoly
K,t
(X).
Proof. It is easy to see that I(t) ▹ K[X] and therefore I(t) = (q(X)) for some monic
generator q(X). To see that q(X) is irreducible, suppose that q(X) = q
1
(X)q
2
(X) with
deg q
i
(X) < deg q(X). Now as q
1
(t)q
2
(t) = 0, we must have q
1
(t) = 0 or q
2
(t) = 0, hence
q
1
(X) ∈ I(t) or q
2
(X) ∈ I(t). These possibilities give q(X)  q
1
(X) or q(X)  q
2
(X) and
so deg q(X) deg q
1
(X) or deg q(X) deg q
2
(X), contradicting the above assumption that
deg q
i
(X) < deg q(X).
The irreducible monic polynomial minpoly
K,t
(X) is in I(t) so q(X)  minpoly
K,t
(X) and
therefore q(X) = minpoly
K,t
(X).
The next Lemma will often be useful.
3.13. Lemma. Let L/K be an extension and suppose that u
1
, . . . , u
n
∈ L are algebraic. Then
K(u
1
, . . . , u
n
)/K is a ﬁnite extension.
Proof. Use induction on n together with Proposition 2.8 and Theorem 2.6(ii).
We now come to an important notion for extensions.
3.14. Definition. The extension L/K is algebraic or L is algebraic over K if every element
t ∈ L is algebraic over K.
37
3.15. Proposition. Let L/K be a ﬁnite extension. Then L/K is algebraic.
Proof. Let t ∈ L. Since the Kvector space L is ﬁnite dimensional, when viewed as
elements of this vector space, the powers 1, t, . . . , t
n
, . . . must be linearly dependent over K.
Hence for suitable coeﬃcients c
j
∈ K not all zero and some m 1 we have
c
0
+c
1
t +· · · +c
m
t
m
= 0.
But this means that t is algebraic over K.
3.16. Proposition. Let M/L and L/K be algebraic extensions. Then the extension M/K
is algebraic.
Proof. Let u ∈ M. Then u is algebraic over L, so there is a polynomial
p(X) = p
0
+p
1
X +· · · +p
m
X
m
∈ L[X]
of positive degree with p(u) = 0. By Lemma 3.13, the extension K(p
0
, . . . , p
m
)/K is ﬁnite and
so is K(p
0
, . . . , p
m
, u)/K(p
0
, . . . , p
m
). By Theorem 2.6(ii), K(p
0
, . . . , p
m
, u)/K is ﬁnite, so by
Proposition 3.15, u is algebraic over K.
3.17. Definition. For an extension L/K, let
L
alg
= {t ∈ L : t is algebraic over K} ⊆ L.
3.18. Proposition. For an extension L/K, L
alg
is a subﬁeld containing K and L
alg
/K is
algebraic.
Proof. Clearly K ⊆ L
alg
. We must show that L
alg
L.
Let u, v ∈ L
alg
. Then by Lemma 3.13, K(u, v)/K is a ﬁnite dimensional extension, hence
every element of K(u, v) is algebraic over K. In particular, u + v and uv are in K(u, v) and if
u ̸= 0, u
−1
is also in K(u, v). Therefore u +v, uv and u
−1
are all algebraic over K.
3.19. Example. In the extension C/Q we can consider C
alg
C which is called the subﬁeld
of algebraic numbers. Similarly, in the extension R/Q the subﬁeld
R
alg
= C
alg
∩ R C
consists of all the real algebraic numbers. Elements of C−C
alg
are called transcendental complex
numbers; examples are e and π. The sets C
alg
and R
alg
are both countable, whereas C and R
are uncountable, so there are in fact many more transcendental numbers but it can be hard to
determine whether a given number is transcendental or not. A more usual notation for C
alg
is Q since this is the algebraic closure of Q which will be discussed later. When dealing with
algebraic extensions of Q we will usually work with subﬁelds of Q = C
alg
.
We end this section with a technical result.
3.20. Proposition. Let K(u)/K be a ﬁnite simple extension. Then there are only ﬁnitely
many subextensions F/K K(u)/K.
Proof. Consider the minimal polynomial minpoly
K,u
(X) ∈ K[X]. Now for any subexten
sion F/K K(u)/K we can also consider
minpoly
F,u
(X) = c
0
+c
1
X +· · · +c
k−1
X
k−1
+X
k
∈ F[X],
38
which divides minpoly
K,u
(X) in F[X]. The Unique Factorization Property 1.33 implies that
minpoly
K,u
(X) has only ﬁnitely many monic divisors in K(u)[X], so there are only a ﬁnite
number of possibilities for minpoly
F,u
(X). Now consider F
0
= K(c
0
, c
1
, . . . , c
k−1
), the extension
ﬁeld of K generated by the coeﬃcients of minpoly
F,u
(X). Then F
0
F and so minpoly
F,u
(X) ∈
F
0
[X] is irreducible since it is irreducible in F[X]; hence minpoly
F,u
(X) = minpoly
F
0
,u
(X). We
have
[K(u) : F] = deg minpoly
F,u
(X) = deg minpoly
F
0
,u
(X) = [K(u) : F
0
],
hence F = F
0
.
This shows that there are only ﬁnitely many subextensions F/K K(u)/K, each of which
has the form K(a
0
, a
1
, . . . , a
ℓ−1
), where
a
0
+a
1
X +· · · +a
ℓ−1
X
ℓ−1
+X
ℓ
∈ K(u)[X]
is a factor of minpoly
K,u
(X) in K(u)[X].
3.2. Splitting ﬁelds and Kronecker’s Theorem
We can now answer a basic question. Let K be a ﬁeld and p(X) ∈ K[X] be a polynomial
of positive degree.
3.21. Question. Is there an extension ﬁeld L/K for which p(X) has a root in L?
A stronger version of this question is the following.
3.22. Question. Is there an extension ﬁeld E/K for which p(X) factorizes into linear factors
in E[X]?
3.23. Definition. p(X) ∈ K[X] splits in E/K or over E if it factorizes into linear factors
in E[X].
Of course, if we have such a ﬁeld E then the distinct roots u
1
, . . . , u
k
of p(X) in E generate
a subﬁeld K(u
1
, . . . , u
k
) E which is the smallest subﬁeld of E that answers Question 3.22.
3.24. Definition. Such a minimal extension of K is called a splitting ﬁeld of p(X) over K
and we will sometimes denote it by K(p(X)) or K
p
.
We already know how to answer Question 3.21.
3.25. Theorem (Kronecker’s Theorem: ﬁrst version). Let K be a ﬁeld and p(X) ∈ K[X]
be a polynomial of positive degree. Then there is a ﬁnite extension L/K for which p(X) has a
root in L.
Proof. We begin by factorizing p(X) ∈ K[X] into irreducible monic factors q
j
(X) together
with a constant factor c:
p(X) = cq
1
(X) · · · q
r
(X).
Now for any j we can form the quotient ﬁeld K[x]/(q
j
(X)) which is a ﬁnite dimensional (simple)
extension of K and in which the coset X + (q
j
(X)) satisﬁes the equation
q
j
(X + (q
j
(X))) = 0 + (q
j
(X)).
Hence p(X) has a root in K[x]/(q
j
(X)).
Of course, this construction is only interesting if q
j
(X) to has degree bigger than 1 since a
linear polynomial already has a root in K.
39
To answer Question 3.22 we iterate this construction. Namely, having found one root u
1
in
an extension L
1
/K we discard the linear factor X −u
1
and consider the polynomial
p
1
(X) =
p(X)
X −u
1
∈ L
1
[X].
We can repeat the argument to form a ﬁnite extension of L
1
(and hence of K) containing a
root of p
1
(X) and so on. At each stage we either already have another root in L
1
or we need
to enlarge the ﬁeld to obtain one.
3.26. Theorem (Kronecker’s Theorem: second version). Let K be a ﬁeld and p(X) ∈ K[X]
be a polynomial of positive degree. Then there is a ﬁnite extension E/K which is a splitting
ﬁeld of p(X) over K.
In practise we often have extension ﬁelds ‘lying around in nature’ containing roots and we
can work inside of these. When working over Q (or any other subﬁeld of C) we can always ﬁnd
roots in C by the Fundamental Theorem of Algebra. We then refer to a subﬁeld of C which is
a splitting ﬁeld as the splitting subﬁeld.
3.27. Example. Find a splitting ﬁeld E/Q for p(X) = X
4
−4 over Q and determine [E : Q].
Solution. Notice that
p(X) = (X
2
−2)(X
2
+ 2),
so ﬁrst we adjoin the roots ±
√
2 of (X
2
− 2) to form Q(
√
2, −
√
2) = Q(
√
2) which gives an
extension Q(
√
2)/Q of degree 2.
Next consider the polynomial X
2
+ 2 ∈ Q(
√
2)[X]. The complex roots of X
2
+ 2 are ±
√
2i
and these are not real, so this polynomial is irreducible in Q(
√
2)[X]. Hence we need to consider
Q(
√
2,
√
2i) = Q(
√
2, i) and the extension Q(
√
2, i)/Q(
√
2) which has degree 2.
C
∞
Q(
√
2, i)
adjoin roots of X
2
+ 2 2
Q(
√
2)
adjoin roots of X
2
− 2 2
Q
Thus the splitting subﬁeld of p(X) over Q in C is Q(
√
2, i) and [Q(
√
2, i) : Q] = 4.
40
Of course we could have started by ﬁrst adjoining roots of X
2
+2 and then adjoining roots
of X
2
−2, thus giving the tower
C
∞
Q(
√
2, i)
adjoin roots of X
2
− 2 2
Q(
√
2i)
adjoin roots of X
2
+ 2 2
Q
An important point is that if a splitting ﬁeld exists inside of a given extension ﬁeld F/K, it is
unique as a subﬁeld of F.
3.28. Proposition. Let F/K be an extension ﬁeld and p(X) ∈ K[X]. If E
1
, E
2
F are
splitting subﬁelds for p(X) over K then E
1
= E
2
.
Proof. Let u
1
, . . . , u
k
∈ F be the distinct roots of p(X) in F. By deﬁnition, K(u
1
, . . . , u
k
)
is the smallest subﬁeld containing K and all the u
j
. But K(u
1
, . . . , u
k
) must be contained in
any splitting subﬁeld, so E
1
= K(u
1
, . . . , u
k
) = E
2
.
Since we will frequently encounter quadratic polynomials we record a useful result on roots
of such polynomials. Recall that p(X) = aX
2
+ bX + c ∈ K[X] is quadratic if a ̸= 0 and its
discriminant is
∆ = b
2
−4ac ∈ K.
The proof of the next result is the standard one which works provided 2 has an inverse in K,
i.e., when char K ̸= 2.
3.29. Proposition. Let K be a ﬁeld of characteristic diﬀerent from 2. Then the quadratic
polynomial p(X) = aX
2
+bX +c ∈ K[X] has
• no roots in K if ∆ is not a square in K;
• one root −b/(2a) = −(2a)
−1
b if ∆ = 0;
• two distinct roots
−b +δ
2a
= (2a)
−1
(−b +δ),
−b −δ
2a
= (2a)
−1
(−b −δ),
if ∆ = δ
2
for some nonzero δ ∈ K.
In particular, the splitting ﬁeld of p(X) over K is K if ∆ is a square in K and K(δ) otherwise,
where δ is one of the two square roots of ∆ in some extension of K such as the algebraic closure
K which we will introduce in Section 3.4.
3.30. Example. Find a splitting ﬁeld E/Q for p(X) = X
3
−2 over Q and determine [E : Q].
Solution. By the Eisenstein Test 1.38, p(X) is irreducible over Q. One root of p(X) is
3
√
2 ∈ R so we adjoin this to Q to form an extension Q(
3
√
2)/Q of degree 3. Now
p(X) = (X −
3
√
2)(X
2
+
3
√
2X + (
3
√
2)
2
)
41
and the second factor has the nonreal complex roots
3
√
2 ζ
3
,
3
√
2 ζ
2
3
lying in the extension
Q(
3
√
2, ζ
3
)/Q(
3
√
2) of degree 2. So the splitting subﬁeld of X
3
− 2 in C over Q is Q(
3
√
2, ζ
3
)
with [Q(
3
√
2, ζ
3
) : Q] = 6.
An alternative strategy would have been to adjoin one of the other roots
3
√
2 ζ
3
or
3
√
2 ζ
2
3
ﬁrst. We could also have begun by adjoining ζ
3
to form the extension Q(ζ
3
)/Q, but none of
the roots of p(X) lie in this ﬁeld so the extension Q(
3
√
2, ζ
3
)/Q(ζ
3
) of degree 3 is obtained by
adjoining one and hence all of the roots.
Figure 3.1 shows all the subﬁelds of the extension Q(
3
√
2, ζ
3
)/Q.
C
∞
2
R
∞ Q(
3
√
2, ζ
3
)
Q(
3
√
2)
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
Q(
3
√
2 ζ
3
)
2
.
.
.
.
.
.
.
.
.
.
3
Q(
3
√
2 ζ
2
3
)
2
3
Q(ζ
3
)
3
2
.
.
.
.
.
.
.
.
.
.
.
Q
Figure 3.1. The subﬁelds of Q(
3
√
2, ζ
3
)/Q
3.3. Monomorphisms between extensions
3.31. Definition. For extensions F/K and L/K, let Mono
K
(L, F) denote the set of all
monomorphisms L −→ F which ﬁx the elements of K.
3.32. Remark. We always have Aut
K
(F) ⊆ Mono
K
(F, F) and Mono
K
(F, F) is closed under
composition but is not always a group since elements are not necessarily invertible. If F/K is
ﬁnite, then we do have Mono
K
(F, F) = Aut
K
(F) since every injective Klinear transformation
is surjective and so invertible.
We will also use the following notation.
3.33. Definition. Let F/K be an extension and p(X) ∈ K[X]. Set
Roots(p, F) = {u ∈ F : p(u) = 0},
the set of roots of p(X) in F. This is always a ﬁnite set which may of course be empty, which
happens precisely when p(X) has no root in F.
42
Suppose that p(X) ∈ K[X] is an irreducible polynomial which we might as well assume is
monic, and let F/K be an extension. Then if t ∈ F is a root of p(X), the evaluation homomor
phism ε
t
: K[X] −→ F factors through the quotient monomorphism ¯ε
t
: K[X]/(p(X)) −→ F
whose image is K(t) F. Of course, there is one such monomorphism for each root of p(X) in
F. If we ﬁx one such root t
0
and identify K[X]/(p(X)) with K(t
0
) via ¯ε
t
0
, then each root of
p(X) in F gives rise to a monomorphism φ
t
= ¯ε
t
◦ ¯ε
−1
t
0
: K(t
0
) −→ F for which φ
t
(t
0
) = t.
K(t
0
)
φ
t
=ε
t
◦ε
−1
t
0
K[X]/(p(X))
ε
t
0
∼
=
ε
t
F
Notice that if φ: K[X]/(p(X)) −→ F is any homomorphism extending the identity function
on K, then the coset X + (p(X)) must be sent by φ to a root of p(X) in F, hence every such
homomorphism arises this way. This discussion is summarized in the following result.
3.34. Proposition. Let F/K be a ﬁeld extension. Let p(X) ∈ K[X] be an irreducible
polynomial with t
0
∈ F be a root of p(X). Then there is a bijection
Roots(p, F) ←→ Mono
K
(K(t
0
), F)
given by t ←→ φ
t
, where φ
t
: K(t
0
) −→ F has the eﬀect φ
t
(t
0
) = t.
3.35. Example. Show that Mono
Q
(Q(
√
2), C) has two elements.
Solution. We have Q(
√
2)
∼
= Q[X]/(X
2
− 2) where X
2
− 2 is irreducible over Q. Hence
the Qmonomorphisms we want send
√
2 to ±
√
2 which are the complex roots of X
2
− 2. In
fact both possibilities occur, giving monomorphisms id, α: Q(
√
2) −→C, where
α(a +b
√
2) = a −b
√
2.
We can replace C by Q(
√
2) to obtain
Mono
Q
(Q(
√
2), C) = Mono
Q
(Q(
√
2), Q(
√
2)) = Aut
Q
(Q(
√
2)).
We will see that this is not always true.
3.36. Example. Show that Mono
Q
(Q(
3
√
2), C) has 3 elements but Mono
Q
(Q(
3
√
2), Q(
3
√
2))
contains only the identity function.
Solution. Here minpoly
Q,
3
√
2
(X) = X
3
− 2 and there are 3 complex roots
3
√
2,
3
√
2 ζ
3
,
3
√
2 ζ
2
3
. As two of these roots are not real, Mono
Q
(Q(
3
√
2), Q(
3
√
2)) contains only the identity
since Q(
3
√
2) R.
Each of the above roots corresponds to one of the subﬁelds Q(
3
√
2), Q(
3
√
2 ζ
3
) or Q(
3
√
2 ζ
2
3
)
of C and there are 3 monomorphisms α
0
, α
1
, α
2
: Q(
3
√
2) −→C given by
α
0
(a +b
3
√
2 +c(
3
√
2)
2
) = a +b
3
√
2 +c(
3
√
2)
2
,
α
1
(a +b
3
√
2 +c(
3
√
2)
2
) = a +b
3
√
2 ζ
3
+c(
3
√
2)
2
ζ
2
3
,
α
2
(a +b
3
√
2 +c(
3
√
2)
2
) = a +b
3
√
2 ζ
2
3
+c(
3
√
2)
2
ζ
3
.
These mappings have images
α
0
Q(
3
√
2) = Q(
3
√
2), α
1
Q(
3
√
2) = Q(
3
√
2 ζ
3
), α
2
Q(
3
√
2) = Q(
3
√
2 ζ
2
3
).
43
3.37. Proposition. Let F/K and L/K be extensions.
(i) For p(X) ∈ K[X], each monomorphism α ∈ Mono
K
(L, F) restricts to a function
α
p
: Roots(p, L) −→ Roots(p, F) which is an injection.
(ii) If α ∈ Mono
K
(L, L), then α
p
: Roots(p, L) −→ Roots(p, L) is a bijection.
Proof. (i) For u ∈ Roots(p, L) we have
p(α(u)) = α(p(u)) = α(0) = 0,
so α maps Roots(p, L) into Roots(p, F). Since α is an injection its restriction to Roots(p, L) ⊆ L
is also an injection.
(ii) From (i), α
p
: Roots(p, L) −→ Roots(p, L) is an injective function from a ﬁnite set to itself,
hence it is also surjective by the Pigeon Hole Principle. Thus α
p
: Roots(p, L) −→ Roots(p, L)
is a bijection.
Part (ii) says that any automorphism of L/K permutes the set of roots in L of a polynomial
p(X) ∈ K[X]. This gives us a strong hold on the possible automorphisms. In the case of ﬁnite,
or more generally algebraic, extensions it is the key to understanding the automorphism group
and this is a fundamental insight of Galois Theory.
3.38. Example. Determine Mono
Q
(Q(
3
√
2, ζ
3
), C).
Solution. We have already met the extension Q(
3
√
2, ζ
3
)/Q in Example 3.30 and we will
make use of information from there. We build up the list of monomorphisms in stages.
First consider monomorphisms that ﬁx
3
√
2 and hence ﬁx the subﬁeld Q(
3
√
2). These form
the subset
Mono
Q(
3
√
2)
(Q(
3
√
2, ζ
3
), C) ⊆ Mono
Q
(Q(
3
√
2, ζ
3
), C).
We know that Q(
3
√
2, ζ
3
) = Q(
3
√
2)(ζ
3
) and that ζ
3
is a root of the irreducible cyclotomic
polynomial Φ
3
(X) = X
2
+ X + 1 ∈ Q(
3
√
2)[X]. So there are two monomorphisms id, α
0
ﬁxing
Q(
3
√
2), where α
0
has the eﬀect
α
0
:
3
√
2 −→
3
√
2
ζ
3
−→ ζ
2
3
.
Next we consider monomorphisms that send
3
√
2 to
3
√
2 ζ
3
. This time we have 2 distinct ways
to extend to elements of Mono
Q
(Q(
3
√
2, ζ
3
), Q(
3
√
2, ζ
3
)) since again we can send ζ
3
to either ζ
3
or ζ
2
3
. The possibilities are
α
1
:
3
√
2 −→
3
√
2 ζ
3
ζ
3
−→ ζ
3
, α
′
1
:
3
√
2 −→
3
√
2 ζ
3
ζ
3
−→ ζ
2
3
.
Finally we consider monomorphisms that send
3
√
2 to
3
√
2 ζ
2
3
. There are again two possibilities
α
2
:
3
√
2 −→
3
√
2 ζ
2
3
ζ
3
−→ ζ
3
, α
′
2
:
3
√
2 −→
3
√
2 ζ
2
3
ζ
3
−→ ζ
2
3
.
These are all 6 of the required monomorphisms. It is also the case here that
Mono
Q
(Q(
3
√
2, ζ
3
), C) = Mono
Q
(Q(
3
√
2, ζ
3
), Q(
3
√
2, ζ
3
)) = Aut
Q
(Q(
3
√
2, ζ
3
)),
so these form a group. It is a nice exercise to show that Aut
Q
(Q(
3
√
2, ζ
3
))
∼
= S
3
, the symmetric
group on 3 objects. It is also worth remarking that  Aut
Q
(Q(
3
√
2, ζ
3
)) = [Q(
3
√
2, ζ
3
) : Q].
44
We end this section with another useful result.
3.39. Proposition. Let L/K be an extension and α ∈ Mono
K
(L, L). Then α restricts to
an automorphism α
alg
: L
alg
−→ L
alg
.
Proof. Suppose that u ∈ L
alg
, say p(u) = 0 for some p(X) ∈ K[X] of positive degree.
Then
p(α(u)) = α(p(u)) = α(0) = 0,
so α maps L
alg
⊆ L into itself and therefore gives rise to a restriction α
alg
: L
alg
−→ L
alg
which
is also a monomorphism. We must show that α
alg
is a bijection by showing it is surjective.
Let v ∈ L
alg
and suppose that q(v) = 0 for some q(X) ∈ K[X] of positive degree. Now
Roots(q, L) ̸= ∅ since it contains v, and it is also ﬁnite. Then α
q
: Roots(q, L) −→ Roots(q, L)
is a bijection by Proposition 3.37(ii), hence v = α
q
(w) = α(w) for some w ∈ Roots(q, L) ⊆ L
alg
.
This shows that v ∈ imα and so α
alg
is surjective.
3.4. Algebraic closures
An important property of the complex numbers is that C is algebraically closed.
3.40. Theorem (Fundamental Theorem of Algebra for C). Every nonconstant polynomial
p(X) ∈ C[X] has a root in C.
3.41. Corollary. Every nonconstant polynomial p(X) ∈ C[X] has a factorization
p(X) = c(X −u
1
) · · · (X −u
d
),
where c, u
1
, . . . , u
d
∈ C and this is unique apart from the order of the roots u
j
.
It is natural to pose the following question.
3.42. Question. Let K be a ﬁeld. Is there an algebraically closed ﬁeld F containing K?
By taking F
alg
we might as well ask that such a ﬁeld be algebraic over K.
3.43. Definition. Let K be a ﬁeld. An extension F/K is called an algebraic closure of K
if F is algebraic over K and algebraically closed.
3.44. Theorem. Let K be a ﬁeld.
(i) There is an algebraic closure of K.
(ii) Let F
1
and F
2
be algebraic closures of K. Then there is an isomorphism φ: F
1
−→ F
2
which ﬁxes the elements of K.
K
·
F
1
φ
F
2
Hence algebraic closures are essentially unique.
Proof. See [3] for a proof using Zorn’s Lemma (see Axiom 3.48) which is logically equiv
alent to the Axiom of Choice.
45
Because of the uniqueness we usually ﬁx some choice of algebraic closure of K and write K
or K
alg cl
, referring to it as the algebraic closure of K. We are already familiar with the example
C = C. There are some immediate consequences of Theorem 3.44. We will temporarily write
E
1
.
= E
2
to indicate that for extensions E
1
/K and E
2
/K there is an isomorphism E
1
−→ E
2
ﬁxing the elements of K.
3.45. Proposition. Let K be a ﬁeld.
(i) If L/K is an algebraic extension, then L
.
= K.
(ii) If L/K is an extension, then so is L/K and (L)
alg
.
= K.
Proof. (i) By Proposition 3.16, every element of L is algebraic over K. Since L is alge
braically closed it is an algebraic closure of K.
(ii) Every nonconstant polynomial in (L)
alg
[X] has a root in L; indeed, by Proposition 3.16,
all of its roots are in fact algebraic over K since (L)
alg
is algebraic over K. Hence these roots
lie in (L)
alg
, which shows that it is algebraically closed.
For example, we have Q = C
alg
and R = C.
There is a stronger result than Theorem 3.44(ii), the Monomorphism Extension Theorem,
which we will ﬁnd useful. Again the proof uses Zorn’s Lemma which we state below. First we
need some deﬁnitions.
3.46. Definition. A partially ordered set (X, ) consists of a set X and a binary relation
such that whenever x, y, z ∈ X,
• x x;
• if x y and y z then x z;
• if x y and y x then x = y.
(X, ) is totally ordered if for every pair x, y ∈ X, at least one of x y or y x is true.
3.47. Definition. Let (X, ) be a partially ordered set and Y ⊆ X.
• y ∈ X is an upper bound for Y if for every y ∈ Y , y y.
• An element x ∈ X is a maximal element of X if
x y =⇒ y = x.
3.48. Axiom (Zorn’s Lemma). Let (X, ) be a partially ordered set in which every totally
ordered subset has an upper bound. Then X has a maximal element.
3.49. Theorem (Monomorphism Extension Theorem). Let M/K be an algebraic extension
and L/K M/K. Suppose that φ
0
: L −→ K is a monomorphism ﬁxing the elements of K.
46
Then there is an extension of φ
0
to a monomorphism φ: M −→ K.
K
M
φ
L
φ
0
K
=
K
Proof. We consider the set X consisting of all pairs (F, θ), where F/L M/L and
θ: F −→ K extends φ
0
. We order X using the relation for which (F
1
, θ
1
) (F
2
, θ
2
) whenever
F
1
F
2
and θ
2
extends θ
1
. Then (X, ) is a partially ordered set.
Suppose that Y ⊆ X is a totally ordered subset. Let
¯
F =
∪
(F,θ)∈Y
F.
Then
¯
F/L M/L. Also there is a function
¯
θ:
¯
F −→ K deﬁned by
¯
θ(u) = θ(u)
whenever u ∈ F for (F, θ) ∈ Y . It is straightforward to check that if u ∈ F
′
for (F
′
, θ
′
) ∈ Y
then
θ
′
(u) = θ(u),
so
¯
θ is welldeﬁned. Then for every (F, θ) ∈ Y we have (F, θ) (
¯
F,
¯
θ), so (
¯
F,
¯
θ) is an upper
bound for Y . By Zorn’s Lemma there must be a maximal element of X, (M
0
, θ
0
).
Suppose that M
0
̸= M, so there is an element u ∈ M for which u / ∈ M
0
. Since M is algebraic
over K it is also algebraic over M
0
, hence u is algebraic over M
0
. If
minpoly
M
0
,u
(X) = a
0
+· · · +a
n−1
X
n−1
+X
n
,
then the polynomial
f(X) = θ
0
(a
0
) +· · · +θ
0
(a
n−1
)X
n−1
+X
n
∈ (θ
0
M
0
)[X]
is also irreducible and so it has a root v in K (which is also an algebraic closure of θ
0
M
0
K). The Homomorphism Extension Property 1.22 of the polynomial ring M
0
[X] applied to
the monomorphism θ
0
: M
0
−→ K yields a homomorphism θ
′
0
: M
0
[X] −→ K extending θ
0
and for which θ
′
0
(u) = v. This factors through the quotient ring M
0
[X]/(minpoly
M
0
,u
(X)) to
give a monomorphism θ
′′
0
: M
0
(u) −→ K extending θ
0
. But then (M
0
, θ
0
) (M
0
(u), θ
′′
0
) and
(M
0
, θ
0
) ̸= (M
0
(u), θ
′′
0
), contradicting the maximality of (M
0
, θ
0
). Hence M
0
= M and so we
can take φ = θ
0
.
3.50. Example. Let u ∈ K and suppose that p(X) = minpoly
K,u
(X) ∈ K[X]. Then for
any other root of p(X), v ∈ K say, there is a monomorphism φ
v
: K(u) −→ K with φ
v
(u) = v.
This extends to a monomorphism φ: K −→ K.
3.51. Definition. Let u, v ∈ K. Then v is conjugate to u over K or is a conjugate of u
over K if there is a monomorphism φ: K −→ K ﬁxing K for which v = φ(u).
47
3.52. Lemma. If u, v ∈ K, then v is conjugate to u over K if and only if minpoly
K,u
(v) = 0.
Proof. Suppose that v = φ(u) for some φ ∈ Mono
K
(K, K). If
minpoly
K,u
(X) = a
0
+a
1
X +· · · +a
d−1
X
d−1
+X
d
,
then
a
0
+a
1
u +· · · +a
d−1
u
d−1
+u
d
= 0
and so
a
0
+a
1
v +· · · +a
d−1
v
d−1
+v
d
= φ(a
0
+a
1
u +· · · +a
d−1
u
d−1
+u
d
) = 0.
The converse follows from Example 3.50.
3.5. Multiplicity of roots and separability
Let K be a ﬁeld. Suppose that f(X) ∈ K[X] and u ∈ K is a root of f(X), i.e., f(u) = 0.
Then we can factor f(X) as f(X) = (X −u)f
1
(X) for some f
1
(X) ∈ K[X].
3.53. Definition. If f
1
(u) = 0 then u is a multiple or repeated root of f(X). If f
1
(u) ̸= 0
then u is a simple root of f(X).
We need to understand more clearly when an irreducible polynomial has a multiple root
since this turns out to be important in what follows. Consider the formal derivative on K[X],
i.e., the function ∂ : K[X] −→ K[X] given by
∂(f(X)) = f
′
(X) = a
1
+ 2a
2
X +· · · +da
d
X
d−1
,
where f(X) = a
0
+a
1
X +a
2
X
2
+· · · +a
d
X
d
with a
j
∈ K.
3.54. Proposition. The formal derivative ∂ : K[X] −→ K[X] has the following properties.
(i) ∂ is Klinear.
(ii) ∂ is a derivation, i.e., for f(X), g(X) ∈ K[X],
∂(f(X)g(X)) = ∂(f(X))g(X) +f(X)∂(g(X)).
(iii) If char K = 0, then ker ∂ = K and ∂ is surjective.
(iv) If char K = p > 0, then
ker ∂ = {h(X
p
) : h(X) ∈ K[X]}
and im∂ is spanned by the monomials X
k
with p (k + 1).
Proof. (i) This is routine.
(ii) By Klinearity, it suﬃces to verify this for the case where f(X) = X
r
and g(X) = X
s
with
r, s 0. But then
∂(X
r+s
) = (r +s)X
r+s−1
= rX
r−1
X
s
+sX
r
X
s−1
= ∂(X
r
)X
s
+X
r
∂(X
s
).
(iii) If f(X) = a
0
+a
1
X +a
2
X
2
+· · · +a
d
X
d
then
∂(f(X)) = 0 ⇐⇒ a
1
= 2a
2
= · · · = da
d
= 0.
So ∂(f(X)) = 0 if and only if f(X) = a
0
∈ K. It is also clear that every polynomial g(X) ∈
K[X] has the form g(X) = ∂(f(X) where f(X) is an antiderivative of g(X).
48
(iv) For a monomial X
m
, ∂(X
m
) = mX
m−1
and this is zero if and only if p  m. Using this we
see that
∂(a
0
+a
1
X +a
2
X
2
+· · · +a
d
X
d
) = 0 ⇐⇒ a
m
= 0 whenever p m.
Also, im∂ is spanned by the monomials X
k
for which ∂(X
k+1
) ̸= 0, which are the ones with
p (k + 1).
We now apply the formal derivative to detect multiple roots.
3.55. Proposition. Let f(X) ∈ K[X] have a root u ∈ L for some extension L/K. Then u
is a multiple root of f(X) if and only if f(X) and f
′
(X) have a common factor of positive
degree in K[X] which vanishes at u.
Proof. Working in L[X], let f(X) = (X −u)f
1
(X). Then
f
′
(X) = f
1
(X) + (X −u)f
′
1
(X),
so f
′
(u) = f
1
(u). Hence u is a multiple root if and only if f(X) and f
′
(X) have a common
factor in L[X] (and hence in K[X] by Proposition 3.12) and which vanishes at u.
3.56. Corollary. If f(X) is irreducible in K[X] then a root u is a multiple root if and
only if f
′
(X) = 0. In particular, this can only happen if char K > 0.
3.57. Corollary. If char K = 0 and f(X) is irreducible in K[X], then every root of f(X)
is simple.
3.58. Example. For n 1, show that each of the roots of f(X) = X
n
−1 in C is simple.
Solution. We have f
′
(X) = ∂(X
n
−1) = nX
n−1
, so for any root ζ of f(X),
f
′
(ζ) = nζ
n−1
̸= 0.
3.59. Example. Show that 2i is a multiple root of f(X) = X
4
+ 8X
2
+ 16.
Solution. We have f
′
(X) = 4X
3
+16X. Using Long Division and the Euclidean Algorithm
we ﬁnd that gcd(f(X), f
′
(X)) = X
2
+4, where 2i is also a root of X
2
+4. Hence 2i is a multiple
root of f(X). In fact, X
4
+ 8X
2
+ 16 = (X
2
+ 4)
2
, so this is obvious.
3.60. Example. Let p > 0 be a prime and suppose that L/F
p
is an extension. Show that
each of the roots of f(X) = X
p
−1 in L is multiple.
Solution. We have f
′
(X) = ∂(X
p
− 1) = pX
p−1
= 0, so if ζ is any root of f(X) then
f
′
(ζ) = 0. Later we will see that 1 is the only root of X
p
−1.
3.61. Definition. An irreducible polynomial p(X) ∈ K[X] is separable over K if every
root of p(X) in an extension L/K is simple. By Corollary 3.56, this is equivalent to requiring
that p
′
(X) ̸= 0. If u ∈ L is a multiple root of p(X), then the multiplicity of u in p(X) is the
maximum m such that p(X) = (X −u)
m
q(X) for some q(X) ∈ L[X].
3.62. Proposition. Let K be a ﬁeld and let K be an algebraic closure. If the irreducible
polynomial p(X) ∈ K[X] has distinct roots u
1
, . . . , u
k
∈ K, then the multiplicities of the u
j
are
equal. Hence in K[X],
p(X) = c(X −u
1
)
m
· · · (X −u
k
)
m
,
where c ∈ K and m 1.
49
Proof. Let u ∈ K be a root of p(X) and suppose that it has multiplicity m, so we can
write p(X) = (X −u)
m
p
1
(X) where p
1
(X) ∈ K(u)[X] and p
1
(u) ̸= 0.
Now let v ∈ K be any other root of p(X). By Proposition 3.34, there is a monomorphism
φ
v
: K(u) −→ K for which φ
v
(u) = v. When p(X) is viewed as an element of K(u)[X], the
coeﬃcients of p(X) are ﬁxed by φ
v
. Then
φ
v
((X −u)
m
p
1
(X)) = (X −u)
m
p
1
(X),
and so
(X −v)
m
¯ p
1
(X) = (X −u)
m
p
1
(X),
where ¯ p
1
(X) ∈ K[X] is obtained applying φ
v
to the coeﬃcients of p
1
(X). Now by Corollary 1.34,
(X − v)
m
must divide p
1
(X) in K[X], and therefore the multiplicity of v must be at least m.
Interchanging the rˆoles of u and v we ﬁnd that the multiplicities of u and v are in fact equal.
3.63. Corollary. Let K be a ﬁeld and let K be an algebraic closure. If the irreducible
polynomial p(X) ∈ K[X] has distinct roots u
1
, . . . , u
k
∈ K which are all simple then in K[X],
p(X) = c(X −u
1
) · · · (X −u
k
),
where c ∈ K and k = deg p(X).
3.64. Corollary. Let K be a ﬁeld and let u ∈ K. Then the number of distinct conjugates
of u is
deg minpoly
K,u
(X)
m
,
where m is the multiplicity of u in minpoly
K,u
(X).
3.65. Definition. An algebraic element u ∈ L in an extension L/K is separable if its
minimal polynomial minpoly
K,u
(X) ∈ K[X] is separable.
3.66. Definition. An algebraic extension L/K is called separable if every element of L is
separable over K.
3.67. Example. An algebraic extension L/K of a ﬁeld of characteristic 0 is separable by
Corollary 3.57.
3.68. Definition. Let L/K be a ﬁnite extension. The separable degree of L over K is
(L : K) =  Mono
K
(L, K).
3.69. Lemma. For a ﬁnite simple extension K(u)/K,
(K(u) : K) =  Roots(minpoly
K,u
, K).
If K(u)/K is separable, then [K(u) : K] = (K(u) : K).
Proof. This follows from Proposition 3.34 applied to the case L = K.
Any ﬁnite extension L/K can be built up from a succession of simple extensions
(3.1) K(u
1
)/K, K(u
1
, u
2
)/K(u
1
), · · · , L = K(u
1
, . . . , u
k
)/K(u
1
, . . . , u
k−1
).
So we can use the following to compute (L : K) = (K(u
1
, . . . , u
k
) : K).
50
3.70. Proposition. Let L/K and M/L be ﬁnite extensions. Then
(M : K) = (M : L)(L : K).
Proof. For α ∈ Mono
K
(M, K) let α
L
∈ Mono
K
(L, K) be its restriction to L. By the
Monomorphism Extension Theorem 3.49, each element of Mono
K
(L, K) extends to a monomor
phism M −→ K, so every element β ∈ Mono
K
(L, K) has the form β = α
L
for some α ∈
Mono
K
(M, K). Since (L : K) =  Mono
K
(L, K), we need to show that the number of such α
is always (M : L) =  Mono
L
(M, K).
So given β ∈ Mono
K
(L, K), choose any extension to a monomorphism
¯
β: K −→ K; by
Proposition 3.39,
¯
β is an automorphism. Of course, restricting to M K we obtain a monomor
phism M −→ K. Now for any extension β
′
: M −→ K of β we can form the composition
¯
β
−1
◦ β
′
: M −→ K; notice that if u ∈ L, then
¯
β
−1
◦ β
′
(u) =
¯
β
−1
(β(u)) = u,
hence
¯
β
−1
◦ β
′
∈ Mono
L
(M, K). Conversely, each γ ∈ Mono
L
(M, K) gives rise to a monomor
phism
¯
β ◦ γ : M −→ K which extends β. In eﬀect, this shows that there is a bijection
¸
extensions of β to monomorphism a M −→ K
¸
←→ Mono
L
(M, K),
so (M : L) =  Mono
L
(M, K) agrees with the number of extensions of β to a monomorphism
M −→ K. Therefore we have the desired formula (M : K) = (M : L)(L : K).
3.71. Corollary. Let L/K be a ﬁnite extension. Then (L : K)  [L : K].
Proof. If L/K is a simple extension then by Propositions 3.62 and 3.34 we know that this
is true. The general result follows by building up L/K as a sequence of simple extensions as
in (3.1) and then using Theorem 2.6(ii) which gives
[L : K] = [K(u
1
) : K] [K(u
1
, u
2
) : K(u
1
)] · · · [K(u
1
, . . . , u
k
) : K(u
1
, . . . , u
k−1
)].
For each k, (K(u
1
, . . . , u
k
) : K(u
1
, . . . , u
k−1
)) divides [K(u
1
, . . . , u
k
) : K(u
1
, . . . , u
k−1
)], so the
desired result follows.
3.72. Proposition. Let L/K be a ﬁnite extension. Then L/K is separable if and only if
(L : K) = [L : K].
Proof. Suppose that L/K is separable. If K E L, then for any u ∈ L, u is alge
braic over E, and in the polynomial ring E[X] we have minpoly
E,u
(X)  minpoly
K,u
(X). As
minpoly
K,u
(X) is separable, so is minpoly
E,u
(X), and therefore L/E is separable. Clearly E/K
is also separable. We have (L : K) = (L : E) (E : K) and [L : K] = [L : E] [E : K], so to
verify that (L : K) = [L : K] it suﬃces to show that (L : E) = [L : E] and (E : K) = [E : K].
Expressing L/K in terms of a sequence of simple extensions as in (3.1), we have
(L : K) = (K(u
1
) : K) · · · (L : K(u
1
, . . . , u
k−1
)),
[L : K] = [K(u
1
) : K] · · · [L : K(u
1
, . . . , u
k−1
)].
Now we can apply Lemma 3.69 to each of these intermediate separable simple extensions to
obtain (L : K) = [L : K].
For the converse, suppose that (L : K) = [L : K]. We must show that for each u ∈ L, u is
separable. For the extensions K(u)/K and L/K(u) we have (L : K) = (L : K(u)) (K(u) : K)
51
and [L : K] = [L : K(u)] [K(u) : K]. By Corollary 3.71, there are some positive integers r, s for
which [L : K(u)] = r(L : K(u)) and [K(u) : K] = s(K(u) : K). Hence
(L : K(u))(K(u) : K) = rs(L : K(u))(K(u) : K),
which can only happen if r = s = 1. Thus (K(u) : K) = [K(u) : K] and so u is separable.
3.73. Proposition. Let L/K and M/L be ﬁnite extensions. Then M/K is separable if and
only if L/K and M/L are separable.
Proof. If M/K is separable then [M : K] = (M : K) and so by Proposition 3.70,
[M : L][L : K] = (M : L)(L : K).
This can only happen if [M : L] = (M : L) and [L : K] = (L : K), since (M : L) [M : L] and
(L : K) [L : K]. By Proposition 3.72 this implies that L/K and M/L are separable.
Conversely, if L/K and M/L are separable then [M : L] = (M : L) and [L : K] = (L : K),
hence
[M : K] = [M : L][L : K] = (M : L)(L : K) = (M : K).
Therefore M/K is separable.
3.6. The Primitive Element Theorem
3.74. Definition. For a ﬁnite simple extension L/K, an element u ∈ L is called a primitive
element for the extension if L = K(u).
3.75. Theorem (Primitive Element Theorem). Let L/K be a ﬁnite separable extension.
Then L has a primitive element.
Proof. The case where K is a ﬁnite ﬁeld will be dealt with in Proposition 5.16. So we will
assume that K is inﬁnite.
Since L is built up from a sequence of simple extensions it suﬃces to consider the case
L = K(u, v). Let p(X), q(X) ∈ K[X] be the minimal polynomials of u and v over K. Suppose
that the distinct roots of p(X) in K are u = u
1
, . . . , u
r
, while the distinct roots of q(X) are
v = v
1
, . . . , v
s
. By the separability assumption, r = deg p(X) and s = deg q(X).
Since K is inﬁnite, we can choose an element t ∈ K for which
t ̸=
u −u
i
v
j
−v
whenever j ̸= 1. Then taking w = u+tv ∈ L, we ﬁnd that w ̸= u
i
+tv
j
whenever j ̸= 1. Deﬁne
the polynomial (of degree r)
h(X) = p(w −tX) ∈ K(w)[X] ⊆ L[X].
Then h(v) = p(u) = 0, but h(v
j
) ̸= p(u
i
) = 0 for any j ̸= 1 by construction of t, so none of the
other v
j
is a zero of h(X).
Now since the polynomials h(X), q(X) ∈ K(w)[X] have exactly one common root in K,
namely v, by separability their greatest common divisor in K(w)[X] is a linear polynomial which
must be X −v, hence v ∈ K(w) and so u = w −tv ∈ K(w). This shows that K(u, v) K(w)
and therefore K(w) = K(u, v).
52
3.76. Corollary. Let L/K be a ﬁnite extension of a ﬁeld of characteristic 0. Then L has
a primitive element.
Proof. Since Q K, K is inﬁnite and by Example 3.67 L/K is separable.
To ﬁnd a primitive element we can always use the method suggested by the proof of Theo
rem 3.75, however a ‘try it and see’ approach will often be suﬃcient.
3.77. Example. Find a primitive element for the extension Q(
√
3, i)/Q.
Solution. Consider
√
3 + i. Then working over the subﬁeld Q(
√
3) Q(
√
3, i) we ﬁnd
that i / ∈ Q(
√
3) R and
(X −(
√
3 +i))(X −(
√
3 −i)) = X
2
−2
√
3X + 4 ∈ Q(
√
3)[X],
hence
X
2
−2
√
3X + 4 = minpoly
Q(
√
3),
√
3+i
(X).
Now taking
(X
2
−2
√
3X + 4)(X
2
+ 2
√
3X + 4) = X
4
−4X
2
+ 16 ∈ Q[X],
we see that minpoly
Q,
√
3+i
(X)  (X
4
−4X
2
+ 16) in Q[X]. Notice that
(
√
3 +i)
−1
=
(
√
3 −i)
(
√
3 +i)(
√
3 −i)
=
(
√
3 −i)
3 + 1
=
1
4
(
√
3 −i) ∈ Q(
√
3 +i),
since (
√
3 +i)
−1
∈ Q(
√
3 +i). Hence
√
3 =
1
2
((
√
3 +i) + (
√
3 −i)), i =
1
2
((
√
3 +i) −(
√
3 −i)),
are both in Q(
√
3 +i), showing that Q(
√
3, i) Q(
√
3 +i) and so Q(
√
3, i) = Q(
√
3 +i). Thus
we must have deg minpoly
Q,
√
3+i
(X) = 4, and so minpoly
Q,
√
3+i
(X) = X
4
−4X
2
+ 16.
There is a general phenomenon illustrated by Example 3.77.
3.78. Proposition. Let u ∈ K be separable over K. Then
minpoly
K,u
(X) = (X −α
1
(u)) · · · (X −α
d
(u)),
where α
1
, . . . , α
d
are the elements of Mono
K
(K(u), K). In particular, the polynomial
(X −α
1
(u)) · · · (X −α
d
(u)) ∈ K[X]
is in K[X] and is irreducible therein.
Proof. Since K(u) is separable then by Lemma 3.52,
d = deg minpoly
K,u
(X) = [K(u) : K] = (K(u) : K).
In Example 3.77 we have
[Q(
√
3, i) : Q] = [Q(
√
3, i) : Q(
√
3)][Q(
√
3) : Q] = 2 · 2 = 4.
There are four monomorphisms α
k
: Q(
√
3, i) −→Q(
√
3, i) given by
α
1
= id, α
2
=
√
3 −→
√
3
i −→ −i
, α
3
=
√
3 −→ −
√
3
i −→ i
, α
4
=
√
3 −→ −
√
3
i −→ −i
.
53
Then
α
2
(
√
3 +i) = (
√
3 −i), α
3
(
√
3 +i) = (−
√
3 +i), α
4
(
√
3 +i) = (−
√
3 −i),
so
(X −
√
3 −i)(X −
√
3 +i)(X +
√
3 −i)(X +
√
3 +i) = X
4
−4X
2
+ 16 ∈ Q[X].
Hence this polynomial is irreducible. So we have [Q(
√
3+i) : Q] = 4 and Q(
√
3+i) = Q(
√
3, i).
3.7. Normal extensions and splitting ﬁelds
Let K be an algebraic closure for the ﬁeld K and let E/K K/K be a ﬁnite extension. If
φ ∈ Mono
K
(E, K), then by Remark 3.32, φE = E if and only if φE E.
3.79. Definition. E/K is normal if φE = E for every φ ∈ Mono
K
(E, K).
3.80. Remark. If E/K is a normal extension then whenever an irreducible polynomial
p(X) ∈ K[X] has a root in E, it splits in E since by Lemma 3.52 each pair of roots of p(X) is
conjugate over K and one can be mapped to the other by a monomorphism K −→ K which
must map E into itself.
3.81. Theorem. A ﬁnite extension E/K is normal if and only if it is a splitting ﬁeld over
K for some polynomial f(X) ∈ K[X].
Proof. Suppose that E/K is normal. Then there is a sequence of extensions
K K(u
1
) K(u
1
, u
2
) · · · K(u
1
, . . . , u
n
) = E
Construct a polynomial by taking
f(X) = minpoly
K,u
1
(X) minpoly
K,u
2
(X) · · · minpoly
K,u
n
(X).
Then by Remark 3.80, f(X) splits in E. Also, E is generated by some of the roots of f(X).
Hence E is a splitting ﬁeld for f(X) over K.
Now suppose that E is a splitting ﬁeld for g(X) ∈ K[X], so that E = K(v
1
, . . . , v
k
), where
v
1
, . . . , v
k
are the distinct roots of g(X) in E. Now any monomorphism θ ∈ Mono
K
(E, K) must
map these roots to θ(v
1
), . . . , θ(v
k
) which are also roots of g(X) and therefore lie in E (see
Proposition 3.34). Since θ permutes the roots v
j
, we have
θE = θK(v
1
, . . . , v
k
) = K(θ(v
1
), . . . , θ(v
k
)) = K(v
1
, . . . , v
k
) = E.
3.82. Corollary. Let E/L and L/K be ﬁnite extensions. If E/K is normal then E/L is
normal.
Proof. If E is the splitting ﬁeld of a polynomial f(X) ∈ K[X] over K, then E is the
splitting ﬁeld of f(X) over L.
These result makes it easy to recognize a normal extension since it is suﬃcient to describe it
as a splitting ﬁeld for some polynomial over K. In Chapter 4 we will see that separable normal
extensions play a central rˆole in Galois Theory, indeed these are known as Galois extensions.
54
Exercises on Chapter 3
3.1. Prove Proposition 3.2.
3.2. Finding splitting subﬁelds E C over Q and determine [E : Q] for each of the following
polynomials.
p
1
(X) = X
4
−X
2
+1, p
2
(X) = X
6
−2, p
3
(X) = X
4
+2, p
4
(X) = X
4
+5X
3
+10X
2
+10X+5.
[Hint: for p
4
(X), consider p
4
(Y −1) ∈ Q[Y ].]
3.3. Prove that Aut
Q
(Q(
3
√
2, ζ
3
))
∼
= S
3
, the symmetric group on 3 elements, as claimed in the
solution of Example 3.38. [Hint: work out the eﬀect of each automorphism on the three roots of
the polynomial X
3
−2.]
3.4. Let k be a ﬁeld of characteristic char k = p > 0 and k(T) be the ﬁeld of rational functions
in T over k. Show that the polynomial g(X) = X
p
− T ∈ k(T)[X] is irreducible and has a
multiple root in k(T). How does g(X) factor in k(T)[X]?
3.5. Find primitive elements for the extensions Q(
√
5,
√
10)/Q, Q(
√
2, i)/Q, Q(
√
3, i)/Q,
Q(
4
√
3, i)/Q, in each case ﬁnding it minimal polynomial over Q. [Hint: look for elements of
high degree over Q, or use the method of proof of Theorem 3.75.]
3.6. Prove the following converse of Proposition 3.20:
Let L/K be a ﬁnite extension. If there are only ﬁnitely many subextensions F/K L/K,
then L/K is simple, i.e., L = K(w) for some w ∈ L.
[Hint: First deal with the case where L = K(u, v), then use induction on n to prove the general
case L = K(u
1
, . . . , u
n
).]
3.7. Let K be a ﬁeld. Show that every quadratic (i.e., of degree 2) extension E/K is normal.
Is such an extension always separable?
3.8. Let f(X) ∈ Q[X] be an irreducible polynomial of odd degree greater than 1 and having
only one real root u ∈ R. Show that Q(u)/Q is not a normal extension.
55
CHAPTER 4
Galois extensions and the Galois Correspondence
In this Chapter we will study the structure of Galois extensions and their associated Galois
groups, in particular we will explain how these are related through the Galois Correspondence.
Throughout the chapter, let K be a ﬁeld.
4.1. Galois extensions
4.1. Definition. A ﬁnite extension E/K is a (ﬁnite) Galois extension if it is both normal
and separable.
From Section 3.5 we know that for such a Galois extension E/K, [E : K] = (E : K) and also
every monomorphism φ ∈ Mono
K
(E, K) maps E into itself, hence restricts to an automorphism
of E which will be denoted φ

E
.
K
E
φ
φ

E
∼
=
E
K
=
K
Also, by the Monomorphism Extension Theorem 3.49, every automorphism α ∈ Aut
K
(E) ex
tends to a monomorphism E −→ K ﬁxing elements of K. So there is a bijection
Mono
K
(E, K) ←→ Aut
K
(E)
and we have
(4.1)  Aut
K
(E) = (E : K) = [E : K].
4.2. Definition. For a ﬁnite Galois extension E/K, the group
Gal(E/K) = Aut
K
(E)
is called the Galois group of the extension or the Galois group of E over K. The elements of
Gal(E/K) are called (Galois) automorphisms of E/K.
Notice that Equation (4.1) implies
(4.2)  Gal(E/K) = (E : K) = [E : K].
We can also reformulate the notion of conjugacy introduced in Deﬁnition 3.51.
4.3. Definition. Let E/K a ﬁnite Galois extension and u, v ∈ E. Then v is conjugate to u
if there is a φ ∈ Gal(E/K) for which v = φ(u); we also say that v is a conjugate of u.
57
It is easy to see that for u, v ∈ K, there is a ﬁnite Galois extension E/K in which v is a
conjugate of u if and only v is a conjugate of u over K in the old sense. Here is a slightly
diﬀerent way to understand this. First notice that every element φ ∈ Aut
K
(K, K) restricts to a
monomorphism E −→ K whose image is contained in E, hence gives rise to an automorphism
φ
E
: E −→ E. Similarly, if F/K is any ﬁnite normal extension with E F, every automorphism
θ: F −→ F restricts to an automorphism θ
F
E
: E −→ E. The proof of the next result is left as
an exercise.
4.4. Proposition. If E/K is a ﬁnite Galois extension, then the function
Aut
K
(K, K) −→ Aut
K
(E, E); φ −→ φ
E
is a surjective group homomorphism. If F/K K/K is any ﬁnite normal extension with E F
then there is a surjective group homomorphism
Aut
K
(F, F) −→ Aut
K
(E, E); θ −→ θ
F
E
.
Furthermore, for φ ∈ Aut
K
(K, K) we have
(φ
F
)
F
E
= φ
E
.
4.2. Working with Galois groups
Let E/K be a ﬁnite Galois extension. Then we know that E is a splitting ﬁeld for some
polynomial over K since E/K is normal. We also know that E is a simple extension of K
since E/K is separable. Hence E is a splitting ﬁeld for the minimal polynomial of any primitive
element for E/K; this minimal polynomial has degree [E : K]. It is often convenient to use these
facts to interpret elements of the Galois group as permutations of the roots of some polynomial
which splits over E.
4.5. Example. Describe the Galois group Gal(Q(
√
2,
√
3)/Q) as a subgroup of the group
of permutations of the roots of (X
2
−2)(X
2
−3) ∈ Q[X].
Solution. We have
[Q(
√
2,
√
3) : Q] = [Q(
√
2,
√
3) : Q(
√
2)] [Q(
√
2) : Q] = 4,
and the following nontrivial elements of the Galois group together with the element identity
α
1
= id:
α
2
=
¸
¸
¸
¸
√
2 −→ −
√
2
−
√
2 −→
√
2
√
3 −→
√
3
−
√
3 −→ −
√
3
¸
, α
3
=
¸
¸
¸
¸
√
2 −→
√
2
−
√
2 −→ −
√
2
√
3 −→ −
√
3
−
√
3 −→
√
3
¸
, α
4
=
¸
¸
¸
¸
√
2 −→ −
√
2
−
√
2 −→
√
2
√
3 −→ −
√
3
−
√
3 −→
√
3
¸
.
Writing the roots in the list
√
2, −
√
2,
√
3, −
√
3 and numbering them from 1 to 4, these auto
morphisms correspond to the following permutations in S
4
expressed in cycle notation:
α
2
←→ (1 2), α
3
←→ (3 4), α
4
←→ (1 2)(3 4).
4.6. Example. Using a primitive element u for the extension, describe the Galois group
Gal(Q(
√
2,
√
3)/Q) as a subgroup of the group of permutations of the roots of minpoly
Q,u
(X) ∈
Q[X].
58
Solution. We have Q(
√
2,
√
3) = Q(
√
2 +
√
3) and the conjugates of u =
√
2 +
√
3 are
±
√
2 ±
√
3. Listing these as
√
2 +
√
3,
√
2 −
√
3, −
√
2 +
√
3, −
√
2 −
√
3,
and after numbering them accordingly, we ﬁnd the correspondences
α
2
←→ (1 3)(2 4), α
3
←→ (1 2)(3 4), α
4
←→ (1 4)(2 3).
Next we summarize the properties of Galois groups that can be deduced from what we
have established so far. Recall that for an extension F/K and a polynomial f(X) ∈ K[X],
Roots(f, F) denotes the set of roots of f(X) in F.
4.7. Recollection. Recall that an action of a group G on a set X is transitive if for every
pair of elements x, y ∈ X, there is an element g ∈ G such that y = gx (so there is only one
orbit); the action is faithful or eﬀective if for every nonidentity element h ∈ G, there is an
element z ∈ X such that hz ̸= z.
4.8. Theorem. Let E/K be a ﬁnite Galois extension. Suppose that E is the splitting ﬁeld
of a separable irreducible polynomial f(X) ∈ K[X] of degree n. Then the following are true.
(i) Gal(E/K) acts transitively and faithfully on Roots(f, E).
(ii) Gal(E/K) can be identiﬁed with a subgroup of the group of permutations of Roots(f, E).
If we order the roots u
1
, . . . , u
n
then Gal(E/K) can be identiﬁed with a subgroup of S
n
.
(iii)  Gal(E/K) divides n! and is divisible by n.
As we have seen in Examples 4.5 and 4.6, in practise it is often easier to use a not necessarily
irreducible polynomial to determine and work with a Galois group.
4.9. Example. The Galois extension Q(ζ
8
)/Q has degree [Q(ζ
8
) : Q] = 4 and it has the
following automorphisms apart from the identity:
α: ζ
8
−→ ζ
3
8
, β: ζ
8
−→ ζ
5
8
, γ : ζ
8
−→ ζ
7
8
.
If we list the roots of the minimal polynomial
minpoly
Q,ζ
(X) = Φ
8
(X) = X
4
+ 1
in the order ζ
8
, ζ
3
8
, ζ
5
8
, ζ
7
8
, we ﬁnd that these automorphisms correspond to the following permu
tations in S
4
:
α ←→ (1 2)(3 4), β ←→ (1 3)(2 4), γ ←→ (1 4)(2 3).
So the Galois group Gal(Q(ζ
8
)/Q) corresponds to
{id, (1 2)(3 4), (1 3)(2 4), (1 4)(2 3)} S
4
.
Noticing that
ζ
8
=
1
√
2
+
1
√
2
i,
we easily ﬁnd that
√
2, i ∈ Q(ζ
8
); hence Q(
√
2, i) Q(ζ
8
). Since [Q(
√
2, i) : Q] = 4, we have
Q(
√
2, i) = Q(ζ
8
). Notice that Q(
√
2, i) is the splitting ﬁeld of f(X) = (X
2
− 2)(X
2
+ 1) over
59
Q. Now list the roots of f(X) in the order
√
2, −
√
2, i, −i, and observe that
α:
¸
¸
¸
¸
√
2 −→ −
√
2
−
√
2 −→
√
2
i −→ −i
−i −→ i
¸
←→ (1 2)(3 4), β:
¸
¸
¸
¸
√
2 −→ −
√
2
−
√
2 −→
√
2
i −→ i
−i −→ −i
¸
←→ (1 2),
γ :
¸
¸
¸
¸
√
2 −→
√
2
−
√
2 −→ −
√
2
i −→ −i
−i −→ i
¸
←→ (3 4).
In this description, the Galois group Gal(Q(ζ
8
)/Q) = Gal(Q(
√
2, i)/Q) corresponds to the
subgroup
{id, (1 2), (3 4), (1 2)(3 4)} S
4
.
While it can be hard to determine Galois groups in general, special arguments can sometimes
be exploited.
4.10. Example. Suppose that f(X) = X
3
+ aX
2
+ bX + c ∈ Q[X] is an irreducible cubic
and that f(X) has only one real root. Then Gal(Q(f(X))/Q)
∼
= S
3
.
Proof. Let u
1
∈ R be the real root of f(X) and let u
2
, u
3
be the remaining complex
roots. Then Q(f(X)) = Q(u
1
, u
2
, u
3
) and in fact [Q(f(X)) : Q] = 6 since [Q(f(X)) : Q]  6
and u
2
/ ∈ Q(u
1
) R. Hence Gal(Q(f(X))/Q) is isomorphic to a subgroup of S
3
and so
Gal(Q(f(X))/Q)
∼
= S
3
since the orders agree. We also have Q(f(X)) ∩ R = Q(u
1
).
The Galois group Gal(Q(f(X))/Q) contains an element of order 3 which corresponds to a
3cycle when viewed as a permutation of the roots u
1
, u
2
, u
3
; we can assume that this is (1 2 3).
It also contains an element of order 2 obtained by restricting complex conjugation to Q(f(X));
this ﬁxes u
1
and interchanges u
2
, u
3
, so it corresponds to the transposition (2 3).
4.11. Remark. Such examples occur when the cubic polynomial f(X) has local maximum
and minimum at real values c
+
and c
−
with f(c
+
), f(c
−
) > 0 or f(c
+
), f(c
−
) < 0. This happens
for example with f(X) = X
3
−3X +3 which has local extrema at ±1 and f(1) = 1, f(−1) = 5.
Given a Galois extension E/K, we will next study subextensions L/K E/K and sub
groups Γ Gal(E/K), focusing on the relationship between objects of these types.
4.3. Subgroups of Galois groups and their ﬁxed ﬁelds
Let E/K a Galois extension and suppose that Γ Gal(E/K). Consider the subset of
elements of E ﬁxed by Γ,
E
Γ
= {u ∈ E : ∀γ ∈ Γ, γ(u) = u}.
4.12. Lemma. E
Γ
E is a subﬁeld of E containing K.
Proof. For u, v ∈ E
Γ
and γ ∈ Γ,
γ(u +v) = γ(u) +γ(v) = u +v, γ(uv) = γ(u)γ(v) = uv.
Also, if u ̸= 0,
γ(u
−1
) = γ(u)
−1
= u
−1
.
60
Finally, if t ∈ K then γ(t) = t, so K E
Γ
.
4.13. Definition. E
Γ
E is the ﬁxed subﬁeld of Γ.
By Proposition 3.73, the extensions E/E
Γ
and E
Γ
/K are separable. E/E
Γ
is also normal,
so this is a Galois extension; we will identify its Galois group. Notice that
[E : E
Γ
] = (E : E
Γ
) =  Gal(E/E
Γ
).
Now each element of Gal(E/E
Γ
) is also an element of Gal(E/K) and Gal(E/E
Γ
) Gal(E/K).
Notice that by deﬁnition Γ Gal(E/E
Γ
), so Lagrange’s Theorem implies that Γ divides
 Gal(E/E
Γ
). In fact we have
4.14. Proposition. For Γ Gal(E/K), we have Gal(E/E
Γ
) = Γ and the equations
[E : E
Γ
] =  Gal(E/E
Γ
) = Γ, [E
Γ
: K] =
 Gal(E/K)
Γ
.
Proof. We know that E/E
Γ
is separable, so by the Primitive Element Theorem 3.75 it is
simple, say E = E
Γ
(u). Now let the distinct elements of Γ be γ
1
= id, γ
2
, . . . , γ
h
, where h = Γ.
Consider the polynomial of degree h
f(X) = (X −u)(X −γ
2
(u)) · · · (X −γ
h
(u)) ∈ E[X].
Notice that f(X) is unchanged by applying any γ
k
to its coeﬃcients since the roots γ
j
(u) are
permuted by γ
k
. Hence, f(X) ∈ E
Γ
[X]. This shows that
[E : E
Γ
] = [E
Γ
(u) : E
Γ
] h = Γ.
Since Γ Gal(E/E
Γ
), we also have
h = Γ  Gal(E/E
Γ
) = [E : E
Γ
].
Combining these two inequalities we obtain
[E : E
Γ
] =  Gal(E/E
Γ
) = Γ = h
and therefore Γ = Gal(E/E
Γ
).
4.4. Subﬁelds of Galois extensions and relative Galois groups
Let E/K a Galois extension and suppose that L/K E/K (i.e., K L E). Then E/L
is also a Galois extension whose Galois group Gal(E/L) is sometimes called the relative Galois
group of the pair of extensions E/K and L/K. The following is immediate.
4.15. Lemma. The relative Galois group of the pair of extensions L/K E/K is a subgroup
of Gal(E/K), i.e., Gal(E/L) Gal(E/K), and its order is  Gal(E/L) = [E : L].
4.16. Proposition. Let L/K E/K. Then L = E
Gal(E/L)
.
Proof. Clearly L E
Gal(E/L)
. Suppose that u ∈ E − L. By Theorem 4.8(i), there is
an automorphism θ ∈ Gal(E/L) such that θ(u) ̸= u, hence u / ∈ E
Gal(E/L)
. This shows that
E
Gal(E/L)
L and therefore E
Gal(E/L)
= L.
We need to understand when Gal(E/L) Gal(E/K) is actually a normal subgroup. The
next result explains the connection between the two uses of the word normal which both ulti
mately derive from their use in Galois theory.
61
4.17. Proposition. Let E/K be a ﬁnite Galois extension and L/K E/K.
(i) The relative Galois group Gal(E/L) of the pair of extensions L/K E/K is a normal
subgroup of Gal(E/K) if and only if L/K is a normal extension.
(ii) If L/K is normal and hence a Galois extension, then there is a group isomorphism
Gal(E/K)/ Gal(E/L)
∼
=
−→ Gal(L/K); αGal(E/L) −→ α

L
.
Proof. (i) Suppose that Gal(E/L) ▹ Gal(E/K), i.e., for all α ∈ Gal(E/L) and β ∈
Gal(E/K), we have βαβ
−1
∈ Gal(E/L). Now if u ∈ L, then for any γ ∈ Gal(E/K) and
α ∈ Gal(E/L), γ(u) ∈ E satisﬁes
αγ(u) = γ(γ
−1
αγ(u)) = γ(u),
since γ
−1
αγ ∈ Gal(E/L); hence γ(u) ∈ E
Gal(E/L)
= L. By the Monomorphism Extension
Theorem 3.49, every monomorphism L −→ K ﬁxing K extends to a monomorphism E −→ K
which must have image E, so the above argument shows that L/K is normal.
Conversely, if L/K is normal, then for every φ ∈ Gal(E/K) and v ∈ L, φ(v) ∈ L, so for
every θ ∈ Gal(E/L), θ(φ(v)) = φ(v) and therefore
φ
−1
θφ(v) = v.
This shows that φ
−1
θφ ∈ Gal(E/L). Hence for every φ ∈ Gal(E/K),
φGal(E/L)φ
−1
= Gal(E/L),
which shows that Gal(E/L) ▹ Gal(E/K).
(ii) If α ∈ Gal(E/K), then αL = L since L/K is normal. Hence we can restrict α to an
automorphism of L,
α

L
: L −→ L; α

L
(u) = α(u).
Then α

L
is the identity function on L if and only if α ∈ Gal(E/L). It is easy to see that the
function
Gal(E/K) −→ Gal(L/K); α −→ α

L
is a group homomorphism whose kernel is Gal(E/L). Thus we obtain an injective homomor
phism
Gal(E/K)/ Gal(E/L) −→ Gal(L/K)
for which
 Gal(E/K)/ Gal(E/L) =
[E : K]
[E : L]
= [L : K] =  Gal(L/K).
Hence this homomorphism is an isomorphism.
4.5. The Galois Correspondence and the Main Theorem of Galois Theory
We are now almost ready to state our central result which describes the Galois Correspon
dence associated with a ﬁnite Galois extension. We will use the following notation. For a ﬁnite
Galois extension E/K, let
S(E/K) = the set of all subgroups of Gal(E/K);
F(E/K) = the set of all subextensions L/K of E/K.
62
Each of these sets is ordered by inclusion. Since every subgroup of a ﬁnite group is a ﬁnite
subset of a ﬁnite set, S(E/K) is also a ﬁnite set. Deﬁne two functions by
Φ
E/K
: F(E/K) −→ S(E/K); Φ
E/K
(L) = Gal(E/L),
Θ
E/K
: S(E/K) −→ F(E/K); Θ
E/K
(Γ) = E
Γ
.
4.18. Theorem (Main Theorem of Galois Theory). Let E/K be a ﬁnite Galois extension.
Then the functions Φ
E/K
and Θ
E/K
are mutually inverse bijections which are orderreversing.
F(E/K)
Φ
E/K
S(E/K)
Θ
E/K
Under this correspondence, normal subextensions of E/K correspond to normal subgroups of
Gal(E/K) and vice versa.
Proof. We know from Proposition 4.16 that for an extension L/K in F(E/K),
Θ
E/K
(Φ
E/K
(L)) = Θ
E/K
(Gal(E/L)) = E
Gal(E/L)
= L.
Also, by Proposition 4.14 for H ∈ S(E/K) we have
Φ
E/K
(Θ
E/K
(Γ)) = Φ
E/K
(E
Γ
) = Gal(E/E
Γ
) = Γ.
This shows that Φ
E/K
and Θ
E/K
are mutually inverse and so are inverse bijections.
Let L
1
/K, L
2
/K ∈ F(E/K) satisfy L
1
/K L
2
/K. Then Gal(E/L
2
) Gal(E/L
1
) since
L
1
⊆ L
2
and so if α ∈ Gal(E/L
2
) then α ﬁxes every element of L
1
. Hence Φ
E/K
(L
2
)
Φ
E/K
(L
1
) and so Φ
E/K
reverses order.
Similarly, if Γ
1
, Γ
2
∈ S(E/K) and Γ
1
Γ
2
, then E
Γ
2
E
Γ
1
since if w ∈ E
Γ
2
then it is ﬁxed
by every element of Γ
1
(as Γ
1
is a subset of Γ
2
). Hence Θ
E/K
reverses order.
There is an immediate consequence of the Main Theorem 4.18 which is closely related to
Proposition 3.20.
4.19. Corollary. Let E/K be a ﬁnite Galois extension. Then there are only ﬁnitely many
subextensions L/K E/K.
Proof. Since the set S(E/K) is ﬁnite, so is F(E/K).
When dealing with a ﬁnite Galois extension E/K, we indicate the subextensions in a diagram
with a line going upwards indicating an inclusion. We can also do this with the subgroups of
the Galois group Gal(E/K) with labels indicating the index of the subgroups. In eﬀect, the
Galois Correspondence inverts these diagrams.
4.20. Example. Figure 4.1 shows the Galois Correspondence for the extension of Exam
ple 3.30.
As noted at the end of Example 3.38, the Galois group here is Gal(Q(
3
√
2, ζ
3
)/Q)
∼
= S
3
. It
is useful to make this isomorphism explicit. First take the 3 roots of the polynomial X
3
−2 for
which E is the splitting ﬁeld over Q; these are
3
√
2,
3
√
2 ζ
3
,
3
√
2 ζ
2
3
which we number in the order
they are listed. Then the monomorphisms id, α
0
, α
1
, α
′
1
, α
2
, α
′
2
extend to automorphisms of E,
each of which permutes these 3 roots in the following ways given by cycle notation:
α
0
= (2 3), α
1
= (1 2 3), α
′
1
= (1 2), α
2
= (1 3 2), α
′
2
= (1 3).
63
E = Q(
3
√
2, ζ
3
)
Q(
3
√
2)
2
3
Q(
3
√
2 ζ
3
)
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
Q(
3
√
2 ζ
2
3
)
2
3
Q(ζ
3
)
3
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Q
Φ
E/K
Gal(E/Q)
Gal(E/Q(ζ
3
))
3
2
Gal(E/Q(
3
√
2))
2
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Gal(E/Q(
3
√
2 ζ
3
))
2
3
Gal(E/Q(
3
√
2 ζ
2
3
))
2
3
{id}
Figure 4.1. The Galois Correspondence for E = Q(
3
√
2, ζ
3
)/Q
We ﬁnd that
Gal(E/Q(ζ
3
)) = {id, α
1
, α
2
}
∼
={id, (1 2 3), (1 3 2)}, Gal(E/Q(
3
√
2)) = {id, α
0
}
∼
={id, (2 3)},
Gal(E/Q(
3
√
2 ζ
3
)) = {id, α
′
2
}
∼
={id, (1 3)}, Gal(E/Q(
3
√
2 ζ
2
3
)) = {id, α
′
1
}
∼
={id, (1 2)}.
Notice that {id, (1 2 3), (1 3 2)} ▹ S
3
and so Q(ζ
3
)/Q is a normal extension. Of course Q(ζ
3
)
is the splitting ﬁeld of X
3
−1 over Q.
4.6. Galois extensions inside the complex numbers and complex conjugation
When working with Galois extensions contained in the complex numbers it is often useful
to make use of complex conjugation as an element of a Galois group. Let E/Q be a ﬁnite Galois
extension with E/Q C/Q. Setting E
R
= R ∩ E, we have Q E
R
E.
4.21. Proposition. Complex conjugation ( ): C −→C restricts to an automorphism of E
over Q, ( )
E/Q
: E −→ E. Furthermore,
(i) ( )
E/Q
agrees with the identity function if and only if E
R
= E.
(ii) If E
R
̸= E, then
⟨
( )
E/Q
⟩
= {id, ( )
E/Q
}
∼
= Z/2,
hence, E
R
= E
⟨( )
E/Q
⟩
and [E : E
R
] = 2.
64
Proof. Let u ∈ E. As E/Q is normal, minpoly
Q,u
(X) ∈ Q[X] splits over E, so all of its
complex roots lie in E. But ( ) permutes the roots of this minimal polynomial. Therefore ( )
maps E into itself.
(i) For z ∈ C, z = z if and only if z ∈ R.
(ii) Here 
⟨
( )
E/Q
⟩
 = 2, and
E
⟨( )
E/Q
⟩
= {u ∈ E : u = u} = E
R
. C
2
∞
R
∞ E
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
E
R
Q
We will usually write ( ) rather than ( )
E/Q
when no confusion seems likely to result.
4.22. Example. Consider the cyclotomic extension Q(ζ
8
)/Q where
ζ
8
= e
πi/4
=
1
√
2
+
1
√
2
i.
From Example 4.9 we know that
Q(ζ
8
) = Q(
√
2, i), [Q(ζ
8
) : Q] = 4,
and we easily see that
Q(ζ
8
)
R
= Q(
√
2).
4.7. Galois groups of even and odd permutations
We have seen that for a monic separable polynomial f(X) ∈ K[X] of degree n, the Galois
group of its splitting ﬁeld E over K can naturally be thought of as a subgroup of the symmetric
group S
n
, where we view the latter as permuting the roots of f(X). It is reasonable to ask when
Gal(E/K) A
n
rather than just Gal(E/K) S
n
.
We ﬁrst recall an interpretation of the sign of a permutation σ ∈ S
n
, sgn σ = ±1. For each
pair i, j with 1 i < j n, exactly one of the inequalities σ(i) < σ(j) or σ(j) < σ(i) must
hold and the ratio (σ(j) − σ(i))/(j − i) is either positive or negative. It is easily veriﬁed that
the righthand side of the following equation must have value ±1 and so
(4.3) sgn σ =
∏
1i<jn
σ(j) −σ(i)
j −i
.
Note that this is sometimes used as the deﬁnition of sgn σ.
65
Suppose that f(X) factorizes over E as
f(X) = (X −u
1
) · · · (X −u
n
) =
n
∏
i=1
(X −u
i
).
Here u
1
, . . . , u
n
∈ E are the roots of f(X); as we have assumed that f(X) is separable, the u
i
are distinct.
4.23. Definition. The discriminant of f(X) is
Discr(f(X)) =
∏
1i<jn
(u
j
−u
i
)
2
∈ E.
Notice that Discr(f(X)) ̸= 0 since u
i
̸= u
j
if i ̸= j.
4.24. Remark. There is an explicit formula for computing Discr(f(X)) is terms of its
coeﬃcients. For polynomials
p(X) = a
0
+a
1
X +· · · +a
m
X
m
, q(X) = b
0
+b
1
X +· · · +b
n
X
n
,
their resultant is the (m+n) ×(m+n) determinant (with n rows of a
i
’s and m rows of b
i
’s)
(4.4) Res(p(X), q(X)) =
a
0
a
1
. . . . . . . a
m
0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
0 a
0
a
1
. . . . . . . a
m
0 . . . . . . . . . . . . . . . . . . . . . . . . . . 0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 a
0
a
1
. . . . . . . a
m
b
0
b
1
. . . . . . . b
n
0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
0 b
0
b
1
. . . . . . . b
n
0 . . . . . . . . . . . . . . . . . . . . . . . . . . 0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0 . . . . . . . . . . . . . . . . . . . . . . . . . . 0 b
0
b
1
. . . . . . . . . . . . b
n
.
Then if f(X) is monic with d = deg f(X),
(4.5) Discr(f(X)) = (−1)
d(d−1)/2
Res(f(X), f
′
(X)).
So for example,
Discr(X
3
+pX +q) = (−1)
3
Res(X
3
+pX +q, 3X
2
+p)
= (−1)
q p 0 1 0
0 q p 0 1
p 0 3 0 0
0 p 0 3 0
0 0 p 0 3
= −4p
3
−27q
2
.
Here are some low degree examples of discriminants obtained with the aid of Maple.
Discr(a
0
+a
1
X +X
2
) = −4a
0
+a
2
1
. n = 2:
Discr(a
0
+a
1
X +a
2
X
2
+X
3
) = −27a
2
0
+ 18a
0
a
1
a
2
+a
2
1
a
2
2
−4a
3
2
a
0
−4a
3
1
. n = 3:
n = 4: Discr(a
0
+a
1
X +a
2
X
2
+a
3
X
3
+X
4
) = 18a
3
a
3
1
a
2
−6a
2
3
a
2
1
a
0
−192a
3
a
1
a
2
0
−27a
4
1
+ 144a
2
a
2
3
a
2
0
+ 144a
0
a
2
1
a
2
+ 256a
3
0
−4a
3
3
a
3
1
−128a
2
2
a
2
0
+ 16a
4
2
a
0
−4a
3
2
a
2
1
+ 18a
3
3
a
1
a
2
a
0
−80a
3
a
1
a
2
2
a
0
−27a
4
3
a
2
0
+a
2
2
a
2
3
a
2
1
−4a
3
2
a
2
3
a
0
.
66
n = 5: Discr(a
0
+a
1
X +a
2
X
2
+a
3
X
3
+a
4
X
4
+X
5
) = 2250a
4
a
2
3
a
3
0
−36a
0
a
3
4
a
3
1
−128a
2
3
a
4
1
+ 2000a
2
0
a
3
a
2
1
−900a
1
a
3
3
a
2
0
−2500a
3
0
a
4
a
1
−50a
2
0
a
2
4
a
2
1
−900a
4
a
3
2
a
2
0
−27a
4
4
a
4
1
−3750a
3
a
2
a
3
0
+ 356a
3
2
a
2
2
a
4
a
1
a
0
+ 560a
3
a
2
2
a
2
4
a
2
0
−2050a
3
a
2
a
2
0
a
4
a
1
−80a
2
3
a
2
a
4
a
1
3
−630a
3
3
a
2
a
4
a
0
2
+ 825a
2
3
a
2
2
a
2
0
+ 16a
3
3
a
3
2
a
0
+ 2000a
2
a
2
4
a
3
0
−6a
2
2
a
2
4
a
3
1
−128a
2
2
a
4
4
a
2
0
+ 16a
4
2
a
3
4
a
0
−4a
3
2
a
3
4
a
2
1
+ 108a
5
3
a
2
0
+ 108a
5
2
a
0
−746a
3
a
2
a
0
a
4
2
a
1
2
−27a
4
2
a
2
1
+ 256a
5
4
a
3
0
−4a
3
3
a
2
2
a
2
1
+ 144a
3
a
2
2
a
3
1
+ 144a
2
4
a
4
1
a
3
+ 3125a
4
0
+ 256a
5
1
−72a
4
3
a
2
a
1
a
0
+ 18a
3
a
2
a
3
4
a
3
1
+ 560a
2
3
a
2
a
0
a
2
1
+ 16a
4
3
a
3
1
+ 18a
3
a
2
3
a
4
a
1
2
−72a
3
a
2
4
a
4
a
0
+ 144a
3
2
a
2
a
4
3
a
0
2
−192a
4
4
a
1
a
3
a
0
2
−630a
3
a
2
3
a
1
a
0
+ 24a
2
3
a
4
2
a
1
a
0
+a
3
2
a
2
2
a
4
2
a
1
2
−6a
4
3
a
1
2
a
3
2
a
0
−80a
3
a
2
2
a
4
3
a
1
a
0
−4a
3
2
a
2
3
a
4
2
a
0
+ 2250a
1
a
2
2
a
2
0
−1600a
3
a
3
4
a
3
0
−192a
4
a
4
1
a
2
−1600a
0
a
3
1
a
2
−4a
3
3
a
3
1
a
2
4
−27a
4
3
a
2
4
a
2
0
+ 1020a
4
2
a
3
2
a
0
2
a
1
+ 18a
3
3
a
2
a
4
2
a
0
a
1
+ 160a
2
a
4
3
a
0
2
a
1
+ 144a
2
a
4
4
a
0
a
1
2
+ 24a
4
a
1
2
a
3
3
a
0
+ 1020a
0
a
4
a
2
2
a
1
2
+ 160a
0
a
4
a
1
3
a
3
.
So for example,
Discr(X
5
+a
4
X
4
+a
0
) = a
3
0
(3125a
0
+ 256a
5
4
), Discr(X
5
+a
1
X +a
0
) = 256a
5
1
+ 3125a
4
0
.
4.25. Proposition. For every σ ∈ Gal(E/K),
σ(Discr(f(X))) = Discr(f(X)).
Hence Discr(f(X)) ∈ E
Gal(E/K)
= K.
Proof. For σ ∈ Gal(E/K) S
n
, we have
σ(Discr(f(X))) =
∏
1i<jn
(u
σ(j)
−u
σ(i)
)
2
=
¸
∏
1i<jn
(u
σ(j)
−u
σ(i)
)
¸
2
.
Now for each pair i, j with i < j,
σ(u
j
−u
i
) = u
σ(j)
−u
σ(i)
,
and by Equation (4.3)
(4.6)
∏
1i<jn
(u
σ(j)
−u
σ(i)
) = sgn σ
∏
1i<jn
(u
j
−u
i
) = (±1)
∏
1i<jn
(u
j
−u
i
).
Hence σ(Discr(f(X))) = Discr(f(X)). Since E
Gal(E/K)
= K, we have Discr(f(X)) ∈ K.
Now let
δ(f(X)) =
∏
1i<jn
(u
j
−u
i
) ∈ E.
Then δ(f(X))
2
= Discr(f(X)), so the square roots of Discr(f(X)) are ±δ(f(X)). Now consider
the eﬀect of σ ∈ Gal(E/K) on δ(f(X)) ∈ E. By Equation (4.6),
σ(δ(f(X))) = sgn σ δ(f(X)) = ±δ(f(X)).
If δ(f(X)) ∈ K, this means that sgn σ = 1. On the other hand, if δ(f(X)) / ∈ K then
K(δ(f(X))) = E
Gal(E/K)∩A
n
.
Of course  Gal(E/K)/ Gal(E/K) ∩ A
n
 = 2.
67
4.26. Proposition. The Galois group Gal(E/K) S
n
is contained in A
n
if and only if
Discr(f(X)) is a square in K.
4.27. Example. For the polynomials of Examples 6.40 and 6.42 we obtain
Discr(X
5
−35X
4
+ 7) = −4611833296875 = −3
3
· 5
6
· 7
4
· 29 · 157,
δ(X
5
−35X
4
+ 7) = ±5
3
· 3 · 7
2
·
√
3 · 29 · 157 i = ±18375
√
13659 i / ∈ Q;
Discr(X
5
+ 20X + 16) = 1024000000 = 2
16
· 5
6
,
δ(X
5
+ 20X + 16) = ±2
8
5
3
∈ Q.
4.8. Kaplansky’s Theorem
In this section we give a detailed account of the Galois theory of irreducible rational poly
nomials f(X) = X
4
+ aX
2
+ b ∈ Q[X]. The following result describes the Galois groups that
occur and the proof introduces some useful computational techniques.
4.28. Theorem (Kaplansky’s Theorem). Let f(X) = X
4
+aX
2
+b ∈ Q[X] be irreducible.
(i) If b is a square in Q then Gal(Q(f(X))/Q)
∼
= Z/2 ×Z/2.
(ii) If b(a
2
−4b) is a square in Q then Gal(Q(f(X))/Q)
∼
= Z/4.
(iii) If neither b nor b(a
2
−4b) is a square in Q then Gal(Q(f(X))/Q)
∼
= D
8
.
Proof. Let g(X) = X
2
+ aX + b ∈ Q[X]. Notice that g(X) must be irreducible since
otherwise f(X) would factorize, hence (a
2
−4b) is not a square in Q. Setting d = (a
2
−4b) ∈ Q
and taking δ to be a square root of d (so δ / ∈ Q), we ﬁnd that the roots of g(X) are (−a±δ)/2 / ∈ Q.
Then the roots of f(X) are ±u, ±v, where
u
2
=
(−a +δ)
2
, v
2
=
(−a −δ)
2
,
so the splitting ﬁeld of f(X) over Q is E = Q(u, v) which contains the quadratic extension
Q(δ)/Q. Since deg f(X) = 4, we also have 4  [E : Q]. In fact, since E is obtained by at most 3
successive quadratic extensions we also have [E : Q]  8.
(i) We have
(uv)
2
= u
2
v
2
=
a
2
−d
4
=
4b
4
= b,
hence uv is a square root of b which is in Q. Setting c = uv ∈ Q, we ﬁnd that v = c/u ∈ Q(u).
This shows that E = Q(u) and we have the following Galois tower.
E = Q(u)
2
Q(δ)
2
Q
In particular [E : Q] = 4 =  Gal(E/Q). Notice that for the Galois extension Q(δ)/Q there
must be a normal subgroup N ▹ Gal(E/Q) with
Q(δ) = E
N
, Gal(Q(δ)/Q) = Gal(E/Q)/N.
68
Hence there is an element σ ∈ Gal(E/Q) for which σ(δ) = −δ. This element must also have
the eﬀects σ(u) = ±v and σ(v) = ±u. Given u we might as well choose v so that σ(u) = v.
There is also an element τ ∈ N for which τ(u) = −u and we also have τ(v) = −v. Notice that
if σ(v) = −u then easy calculation shows that
τσ(v) = στ(v) = u, τσ(δ) = στ(δ) = −δ,
hence we might as assume that σ(v) = u since if necessary we can replace our original choice
by τσ.
We now have
σ(u) =
c
u
, τ(u) = −u, τσ(u) = στ(u) = −
c
u
.
These satisfy
σ
2
= τ
2
= (στ)
2
= id = the identity, στ = τσ.
This shows that
Gal(Q(f(X))/Q) = Gal(E/Q) = {id, σ, τ, στ}
∼
= Z/2 ×Z/2 = the Klein 4group.
(ii) If bd is a square in Q, then
(uvδ)
2
= u
2
v
2
d = bd,
which is a square in Q, so we can write uvδ = c ∈ Q or equivalently v = c/(uδ) ∈ Q(u) since
Q(δ) Q(u). This shows that E = Q(u, v) = Q(u) and again we have a Galois tower
E = Q(u)
2
Q(δ)
2
Q
with [E : Q] = 4 =  Gal(E/Q).
Since Q(δ)/Q is Galois there is an element σ ∈ Gal(E/Q) with σ(δ) = −δ and this has the
eﬀect σ(u) = ±v; given u we might as well choose v so that σ(u) = v. Notice that
σ(v) =
c
σ(uδ)
= −
c
vδ
= −u,
so σ
2
(u) = −u. This shows that
Gal(Q(f(X))/Q) = Gal(E/Q) = {id, σ, σ
2
σ
3
}
∼
= Z/4 = a cyclic group of order 4.
(iii) Suppose that d, b and bd are not squares in Q. By an easy calculation we ﬁnd that (uv)
2
= b,
so uv ∈ E is a square root of b in E. Suppose that uv ∈ Q(δ); then uv = p+qδ for some p, q ∈ Q.
By squaring we obtain
b = (p
2
+q
2
d) + 2pqδ,
and so pq = 0. We cannot have q = 0 since this would imply that b was a square in Q; if
p = 0 then b = q
2
d and so bd = (qd)
2
, implying that bd was a square in Q. Thus we have
Q(uv) ∩ Q(δ) = Q. A similar discussion shows that
Q(uvδ) ∩ Q(δ) = Q = Q(uvδ) ∩ Q(uv).
69
So we have a Galois tower which includes the following subﬁelds.
E = Q(u, v)
Q(uv, δ)
.
.
.
.
.
.
.
.
.
.
Q(δ)
2
Q(uv)
2
Q(uvδ)
2
.
.
.
.
.
.
.
.
.
.
.
.
Q
Choose
α ∈ Gal(E/Q(uv)) Gal(E/Q)
so that α(δ) = −δ. By renaming −v to v if necessary, we may assume that v = α(u) and so
u = α(v). Notice that α
2
= id.
Choose
β ∈ Gal(E/Q(δ)) Gal(E/Q)
with β(uv) = −uv. We must have either β(u) = −u or β(v) = −v, so by interchanging ±δ if
necessary we can assume that β(u) = −u and β(v) = v. Notice that β
2
= id.
Choose
γ ∈ Gal(E/Q(δ, uv)) Gal(E/Q)
so that γ(u) = −u. Then we must have γ(v) = −v since γ(uv) = uv. Notice that γ
2
= id.
Setting σ = αβ we ﬁnd σ(u) = −v and σ(v) = u. Then σ
2
= γ and σ has order 4. Also,
ασα = βσβ = σ
−1
.
The eight elements
id, σ, γ, σ
−1
, α, ασ, αγ, ασ
−1
form a group isomorphic to the dihedral group of order 8, D
8
. Therefore we have
Gal(Q(f(X))/Q) = Gal(E/Q)
∼
= D
8
,
and [E : Q] = 8. The corresponding Galois tower is
E = Q(u, v)
2
Q(uv, δ)
2
.
.
.
.
.
.
.
.
.
.
2
2
Q(δ)
2
Q(uv)
2
Q(uvδ)
2
.
.
.
.
.
.
.
.
.
.
.
.
Q
70
4.29. Example. We have the following Galois groups:
Gal(Q(X
4
+ 1)/Q)
∼
= Z/2 ×Z/2; Gal(Q(X
4
+ 4X
2
+ 2)/Q)
∼
= Z/4;
Gal(Q(X
4
+ 2X
2
+ 2)/Q)
∼
= D
8
.
Exercises on Chapter 4
4.1. If f(X) ∈ K[X] is a separable polynomial, prove that the splitting ﬁeld of f(X) over K
is a ﬁnite Galois extension of K.
4.2. Let K be a ﬁeld for which char K ̸= 2, 3 and suppose that f(X) ∈ K[x] is a cubic
polynomial.
(a) Show that there u, v ∈ K with u ̸= 0 such that f(uX + v) = X
3
+ aX + b for some
a, b ∈ K. If f(X) is monic, deduce that a, b ∈ K; under what conditions is this always
true?
(b) If g(X) = X
3
+ aX + b ∈ K[x] is irreducible and E = K(g(X)) is its splitting ﬁeld
over K, explain why Gal(E/K) is isomorphic to one of the groups S
3
or A
3
.
(c) Continuing with the notation and assumptions of (b), suppose that w
1
, w
2
, w
3
are the
distinct roots of g(X) in E and let
∆ = (w
1
−w
2
)
2
(w
2
−w
3
)
2
(w
1
−w
3
)
2
∈ E.
Show that
∆ = −4b
3
−27a
2
,
and hence ∆ ∈ K. If δ = (w
1
−w
2
)(w
3
−w
3
)(w
1
−w
3
), show that
Gal(E/K)
∼
=
A
3
if δ ∈ K,
S
3
if δ / ∈ K.
[Hint: Consider K(δ) E and the eﬀect on the element δ of even and odd permutations
in Gal(E/K) S
3
.]
4.3. Show that f(X) = X
3
−3X+1 ∈ Q[X] is irreducible over Q, and show that its discriminant
is a square in Q. Prove that the Galois group of f(X) over Q is cyclic.
4.4. This is a revision exercise on ﬁnite groups of small order.
(a) Show that every nonabelian ﬁnite group has order at least 6.
(b) Let D
8
be the dihedral group with the eight elements
ι, α, α
2
, α
3
, β, βα, βα
2
, βα
3
satisfying
α
4
= ι, β
2
= ι, βαβ = α
−1
= α
3
.
Find all the normal subgroups of D
8
.
71
4.5. Use Kaplansky’s Theorem 4.28 to ﬁnd the Galois group of the splitting ﬁeld E of the
polynomial X
4
+ 3 ∈ Q[X] over Q. Determine all the subextensions F E for which F/Q is
Galois.
4.6. Find the Galois groups for each of the following extensions:
Q(X
3
−10)/Q; Q(
√
2)(X
3
−10)/Q(
√
2); Q(
√
3 i)(X
3
−10)/Q(
√
3 i);
Q(
√
23 i)(X
3
−X −1)/Q(
√
23 i); K(X
3
−X −1)/K for K = Q, Q(
√
5), Q(
√
5 i), Q(i).
4.7. Let p > 0 be a prime. Let K be a ﬁeld with char K ̸= p. Suppose that 0 ̸= a ∈ K and
f(X) = X
p
−a ∈ K[X]. Let L/K where L is a splitting ﬁeld for f(X) over K.
(a) Show that f(X) has p distinct roots in L. If u ∈ L is one such root, describe the
remaining roots and show that L contains p distinct pth roots of 1.
(b) Suppose that K contains p distinct pth roots of 1. Show that either f(X) is irreducible
over K or it factors into p distinct linear factors over K.
(c) Suppose that the only pth root of 1 in K is 1. Show that either f(X) is irreducible
over K or it has a root in K.
4.8. Let K be a ﬁeld of characteristic char K = p where p > 0 is a prime. Suppose that
0 ̸= a ∈ K and f(X) = X
p
− a ∈ K[X]. Show that if f(X) has no root in K then it is
irreducible over K.
72
CHAPTER 5
Galois extensions for ﬁelds of positive characteristic
In this chapter we will investigate extensions of ﬁelds of positive characteristic, especially
ﬁnite ﬁelds. A thorough account of ﬁnite ﬁelds and their applications can be found in [6].
Throughout this chapter we will assume that K is a ﬁeld of prime characteristic p =
char K > 0, containing the prime subﬁeld F
p
.
5.1. Finite ﬁelds
If K is a ﬁnite ﬁeld, then K is an F
p
vector space. Our ﬁrst goal is to count the elements
of K. Here is a more general result.
5.1. Lemma. Let F be a ﬁnite ﬁeld with q elements and let V be an Fvector space. Then
dim
F
V < ∞ if and only if V is ﬁnite in which case V  = q
dim
F
V
.
Proof. If d = dim
F
V < ∞, then for a basis v
1
, . . . , v
d
we can express each element v ∈ V
uniquely in the form v = t
1
v
1
+ · · · + t
d
v
d
, where t
1
, . . . , t
d
∈ F. Clearly there are exactly q
d
such expressions, so V  = q
d
.
Conversely, if V is ﬁnite then any basis has ﬁnitely many elements and so dim
F
V < ∞.
5.2. Corollary. Let F be a ﬁnite ﬁeld and E/F an extension. Then E is ﬁnite if and
only if E/F is ﬁnite and then E = F
[E:F]
.
5.3. Corollary. Let K be a ﬁnite ﬁeld. Then K/F
p
is ﬁnite and K = p
[K:F
p
]
.
Our next task is to show that for each power p
d
there is a ﬁnite ﬁeld with p
d
elements. We
start with the algebraic closure F
p
of F
p
and consider the polynomial
Θ
p
d(X) = X
p
d
−X ∈ F
p
[X].
Notice that Θ
′
p
d
(X) = −1, hence by Proposition 3.55 every root of Θ
p
d(X) in F
p
is sim
ple. Therefore by Corollary 1.35 Θ
p
d(X) must have exactly p
d
distinct roots in F
p
, say
0, u
1
, . . . , u
p
d
−1
. Then in F
p
[X] we have
X
p
d
−X = X(X −u
1
) · · · (X −u
p
d
−1
),
and each root is separable over F
p
. Let
F
p
d = {u ∈ F
p
: Θ
p
d(u) = 0} ⊆ F
p
, F
0
p
d
= {u ∈ F
p
d : u ̸= 0}.
Notice that u ∈ F
0
p
d
if and only if u
p
d
−1
= 1.
5.4. Proposition. For each d 1, F
p
d is a ﬁnite subﬁeld of F
p
with p
d
elements and
F
0
p
d
= F
×
p
d
. Furthermore, the extension F
p
d/F
p
is a separable splitting ﬁeld.
73
Proof. If u, v ∈ F
p
d then by the Idiot’s Binomial Theorem 1.10,
(u +v)
p
d
−(u +v) = (u
p
d
+v
p
d
) −(u +v) = (u
p
d
−u) + (v
p
d
−v) = 0,
(uv)
p
d
−uv = u
p
d
v
p
d
−uv = uv −uv = 0.
Furthermore, if u ̸= 0 then u
p
d
−1
= 1 and so u has multiplicative inverse u
p
d
−2
. Hence F
p
d F
p
.
Notice that F
p
F
p
d, so F
p
d/F
p
is a ﬁnite extension. In any ﬁeld the nonzero elements are
always invertible, hence F
0
p
d
= F
×
p
d
.
5.5. Definition. The ﬁnite subﬁeld F
p
d F
p
is called the Galois ﬁeld of order p
d
.
The notation GF(p
d
) is often used in place of F
p
d. Of course, F
p
1 = GF(p
1
) = GF(p) = F
p
and [F
p
d : F
p
] = d.
5.6. Proposition. Let d 1.
(i) F
p
d F
p
is the splitting subﬁeld for each of the polynomials X
p
d
− X and X
p
d
−1
− 1
over F
p
.
(ii) F
p
d F
p
is the unique subﬁeld with p
d
elements.
(iii) If F is any ﬁeld with p
d
elements then there is a monomorphism F −→F
p
with image
F
p
d, hence F
∼
= F
p
d.
Proof. (i) As F
p
d consists of exactly the roots of Θ
p
d(X) in F
p
, it is the splitting subﬁeld.
The nonzero elements of F
p
d are the roots of X
p
d
−1
−1, so F
p
d is also the splitting subﬁeld for
this polynomial.
(ii) Let F F
p
have p
d
elements. Notice that the nonzero elements of F form a group F
×
under multiplication. This group is abelian and has p
d
−1 elements, so by Lagrange’s Theorem,
each element u ∈ F
×
has order dividing p
d
− 1, therefore u
p
d
−1
= 1 and so u
p
d
= u. But this
means every element of F is a root of Θ
p
d(X) and so F F
p
d; equality follows since these
subﬁelds both have p
d
elements.
(iii) Apply the Monomorphism Extension Theorem 3.49 for K = L = F
p
and M = F. By (ii),
the image of the resulting monomorphism must be F
p
d, therefore F
∼
= F
p
d.
It is worth noting the following consequence of this result and the construction of F
p
d.
5.7. Corollary. Let K be a ﬁnite ﬁeld of characteristic p. Then K/F
p
is a ﬁnite Galois
extension.
5.8. Example. Consider the polynomial X
4
−X ∈ F
2
[X]. By inspection, in the ring F
2
[X]
we ﬁnd that
X
4
−X = X
4
+X = X(X
3
+ 1) = X(X + 1)(X
2
+X + 1).
Now X
2
+ X + 1 has no root in F
2
so it must be irreducible in F
2
[X]. Its splitting ﬁeld is a
quadratic extension F
2
(w)/F
2
where w is one of the roots of X
2
+X +1, the other being w+1
since the sum of the roots is the coeﬃcient of X. This tells us that every element of F
4
= F
2
(w)
can be uniquely expressed in the form a +bw with a, b ∈ F
2
. To calculate products we use the
fact that w
2
= w + 1, so for a, b, c, d ∈ F
2
we have
(a +bw)(c +dw) = ac + (ad +bc)w +bdw
2
= (ac +bd) + (ad +bc +bd)w.
74
5.9. Example. Consider the polynomial X
9
− X ∈ F
3
[X]. Let us ﬁnd an irreducible
polynomial of degree 2 in F
3
[X]. Notice that X
2
+1 has no root in F
3
, hence X
2
+1 ∈ F
3
[X] is
irreducible; so if u ∈ F
3
is a root of X
2
+ 1 then F
3
(u)/F
3
has degree 2 and F
3
(u) = F
9
. Every
element of F
9
can be uniquely expressed in the form a + bu with a, b ∈ F
3
. Multiplication is
carried out using the relation u
2
= −1 = 2.
By inspection, in the ring F
3
[X] we ﬁnd that
X
9
−X = X(X
8
−1) = (X
3
−X)(X
2
+ 1)(X
2
+X −1)(X
2
−X −1).
So X
2
+X−1 and X
2
−X−1 are also quadratic irreducibles in F
3
[X]. We can ﬁnd their roots
in F
9
using the quadratic formula since in F
3
we have 2
−1
= (−1)
−1
= −1. The discriminant of
X
2
+X −1 is
1 −4(−1) = 5 = 2 = u
2
,
so its roots are (−1)(−1 ±u) = 1 ±u. Similarly, the discriminant of X
2
−X −1 is
1 −4(−1) = 5 = 2 = u
2
and its roots are (−1)(1 ±u) = −1 ±u. Then we have
F
9
= F
3
(u) = F
3
(1 ±u) = F
3
(−1 ±u).
There are two issues we can now clarify.
5.10. Proposition. Let F
p
m and F
p
n be two Galois ﬁelds of characteristic p. Then F
p
m
F
p
n if and only if m  n.
Proof. If F
p
m F
p
n, then by Corollary 5.2,
p
n
= (p
m
)
[F
p
n:F
p
m]
= p
m[F
p
n:F
p
m]
,
so m  n.
If m  n, write n = km with k 1. Then for u ∈ F
p
m we have u
p
m
= u, so
u
p
n
= u
p
mk
= (u
p
m
)
p
m(k−1)
= u
p
m(k−1)
= · · · = u
p
m
= u.
Hence u ∈ F
p
n and therefore F
p
m F
p
n.
This means that we can think of the Galois ﬁelds F
p
n as ordered by divisibility of n. The dia
gram of subﬁelds for F
p
24 can be seen in Figure 5.1 which shows extensions with no intermediate
subextensions.
5.11. Theorem. The algebraic closure of F
p
is the union of all the Galois ﬁelds of charac
teristic p,
F
p
=
∪
n1
F
p
n.
Furthermore, each element u ∈ F
p
is separable over F
p
.
Proof. Let u ∈ F
p
. Then u is algebraic over F
p
and the extension F
p
(u)/F
p
is ﬁnite. Hence
by Corollary 5.2, F
p
(u) F
p
is a ﬁnite subﬁeld. Proposition 5.10 now implies that F
p
(u) = F
p
n
for some n. The separability statement follows from Corollary 5.7.
We will require a useful fact about Galois ﬁelds.
5.12. Proposition. The group of units F
×
p
d
in F
p
d is cyclic.
75
F
p
24
F
p
8 F
p
12
F
p
4 F
p
6
F
p
2
F
p
3
F
p
Figure 5.1. The subﬁelds of F
p
24
This is a special case of a more general result about arbitrary ﬁelds.
5.13. Proposition. Let K be a ﬁeld. Then every ﬁnite subgroup U K
×
is cyclic.
Proof. Use Corollary 1.35 and Lemma 1.46.
5.14. Definition. w ∈ F
×
p
d
is called a primitive root if it is a primitive (p
d
− 1)th root of
unity, i.e., its order in the group F
×
p
d
is (p
d
−1), hence ⟨w⟩ = F
×
p
d
.
5.15. Remark. Unfortunately the word primitive has two confusingly similar uses in the
context of ﬁnite ﬁelds. Indeed, some authors use the term primitive element for what we have
called a primitive root, but that conﬂicts with our usage, although as we will in the next result,
every primitive root is indeed a primitive element in our sense!
5.16. Proposition. The extension of Galois ﬁelds F
p
nd/F
p
d is simple, i.e., F
p
nd = F
p
d(u)
for some u ∈ F
p
nd.
Proof. By Proposition 5.12, F
p
nd has a primitive root w say. Then every element of
F
p
nd can be expressed as a polynomial in w, so F
p
nd F
p
d(w) F
p
nd. This implies that
F
p
nd = F
p
d(w).
5.17. Remark. This completes the proof of the Primitive Element Theorem 3.75 which we
had previously only established for inﬁnite ﬁelds.
5.18. Example. In Example 5.8 we ﬁnd that F
4
= F
2
(w) has the two primitive roots w
and w + 1.
5.19. Example. In Example 5.9 we have F
9
= F
3
(u) and F
×
9
is cyclic of order 8. Since
φ(8) = 4, there are four primitive roots and these are the roots of the polynomials X
2
+X −1
and X
2
−X −1 which we found to be ±1 ±u.
We record a fact that is very important in Number Theory.
5.20. Proposition. Let p > 0 be an odd prime.
(i) If p ≡ 1 (mod 4), the polynomial X
2
+ 1 ∈ F
p
[X] has two roots in F
p
.
76
(ii) If p ≡ 3 (mod 4) the polynomial X
2
+1 ∈ F
p
[X] is irreducible, so F
p
2
∼
= F
p
[X]/(X
2
+1).
Proof. (i) We have 4  (p −1) = F
×
p
, so if u ∈ F
×
p
is a generator of this cyclic group, the
order of u
F
×
p
/4
is 4, hence this is a root of X
2
+ 1 (the other root is −u
F
×
p
/4
).
(ii) If v ∈ F
p
is a root of X
2
+1 then v has order 4 in F
×
p
. But then 4  (p −1) = F
×
p
, which is
impossible since p −1 ≡ 2 (mod 4).
Here is a generalization of Proposition 5.20.
5.21. Proposition. F
p
d contains a primitive nth root of unity if and only if p
d
≡ 1 (mod n)
and p n.
5.2. Galois groups of ﬁnite ﬁelds and Frobenius mappings
Consider an extension of Galois ﬁelds F
p
nd/F
p
d. By Proposition 5.6(i), Corollary 5.7 and
Proposition 3.73, this extension is Galois and
 Gal(F
p
nd/F
p
d) = [F
p
nd : F
p
d] = n.
We next introduce an important element of the Galois group Gal(F
p
nd/F
p
d).
5.22. Definition. The (relative) Frobenius mapping for the extension F
p
nd/F
p
d is the func
tion F
d
: F
p
nd −→F
p
nd given by F
d
(t) = t
p
d
.
5.23. Proposition. The relative Frobenius mapping F
d
: F
p
nd −→ F
p
nd is an automor
phism of F
p
nd that ﬁxes the elements of F
p
d, so F
d
∈ Gal(F
p
nd/F
p
d). The order of F
d
is n, so
Gal(F
p
nd/F
p
d) = ⟨F
d
⟩, the cyclic group generated by F
d
.
Proof. For u, v ∈ F
p
nd, we have the identities
F
d
(u +v) = (u +v)
p
d
= u
p
d
+v
p
d
, F
d
(uv) = (uv)
p
d
= u
p
d
v
p
d
,
so F
d
is a ring homomorphism. Also, for u ∈ F
p
d we have
F
d
(u) = u
p
d
= u,
so F
d
ﬁxes the elements of F
p
d. To see that F
d
is an automorphism, notice that the composition
power F
n
d
= F
d
◦ · · · ◦ F
d
(with n factors) satisﬁes
F
n
d
(t) = t
p
nd
= t
for all t ∈ F
p
nd, hence F
n
d
= id. Then F
d
is invertible with inverse F
−1
d
= F
n−1
d
. This also shows
that the order of F
d
in the group Aut
F
p
d
(F
p
nd) is at most n. Suppose the order is k with k n;
then every element u ∈ F
p
nd satisﬁes the equation F
k
d
(u) = u which expands to u
p
kd
= u, hence
u ∈ F
p
kd. But this can only be true if k = n.
Frobenius mappings exist on the algebraic closure F
p
. For d 1, consider the function
F
d
: F
p
−→F
p
; F
d
(t) = t
p
d
.
5.24. Proposition. Let d 1.
(i) F
d
: F
p
−→ F
p
is an automorphism of F
p
which ﬁxes the elements of F
p
d. In fact for
u ∈ F
p
, F
d
(u) = u if and only if u ∈ F
p
d.
77
(ii) The restriction of F
d
to the Galois subﬁeld F
p
dn agrees with the relative Frobenius
mapping F
d
: F
p
nd −→F
p
nd.
(ii) If k 1, then F
k
d
= F
kd
. Hence in the automorphism group Aut
F
p
d
(F
p
), F
d
has inﬁnite
order, so Aut
F
p
d
(F
p
) is inﬁnite.
Proof. This is left as an exercise.
The Frobenius mapping F = F
1
is often called the absolute Frobenius mapping since it exists
as an element of each of the groups Aut
F
p
(F
p
) and Aut
F
p
(F
p
n) = Gal(F
p
n/F
p
) for every n 1.
In Gal(F
p
nd/F
p
d) = ⟨F
d
⟩, for each k with k  n there is the cyclic subgroup
⟨
F
k
d
⟩
of order

⟨
F
k
d
⟩
 = n/k.
5.25. Proposition. For k  n, the ﬁxed subﬁeld of
⟨
F
k
d
⟩
in F
p
nd is F
⟨F
k
d
⟩
p
nd
= F
p
dk.
F
p
nd
n/k
F
⟨F
k
d
⟩
p
nd
= F
p
dk
k
F
p
d
Proof. For u ∈ F
p
nd we have F
k
d
(u) = u
p
dk
, hence F
k
d
(u) = u if and only if u ∈ F
p
dk.
Figure 5.2 shows the subgroup diagram corresponding to the lattice of subﬁelds of F
p
24
shown in Figure 5.1.
Gal(F
p
24/F
p
) = ⟨F⟩
∼
= Z/24
⟨
F
2
⟩
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. ⟨
F
3
⟩
⟨
F
4
⟩ ⟨
F
6
⟩
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
⟨
F
8
⟩ ⟨
F
12
⟩
⟨
F
24
⟩
= {id}
Figure 5.2. The subgroups of the Galois groups of F
p
24/F
p
78
5.3. The trace and norm mappings
For an extension of Galois ﬁelds F
p
nd/F
p
d, consider the function T
F
p
nd
/F
p
d
: F
p
nd −→ F
p
nd
deﬁned by
T
F
p
nd
/F
p
d
(u) = u +u
p
d
+u
p
2d
+· · · +u
p
(n−1)d
= u + F
d
(u) + F
2d
(u) +· · · + F
(n−1)d
(u).
Notice that
F
d
(T
F
p
nd
/F
p
d
(u)) = u
p
d
+u
p
2d
+u
p
3d
+· · · +u
p
nd
= u
p
d
+u
p
2d
+u
p
3d
+· · · +u
p
(n−1)d
+u = T
F
p
nd
/F
p
d
(u).
So by Proposition 5.24(i), T
F
p
nd
/F
p
d
(u) ∈ F
p
d. If we modify T
F
p
nd
/F
p
d
to have codomain F
p
d,
we obtain the relative trace
Tr
F
p
nd
/F
p
d
: F
p
nd −→F
p
d; Tr
F
p
nd
/F
p
d
(u) = u +u
p
d
+u
p
2d
+· · · +u
p
(n−1)d
.
5.26. Proposition. The relative trace Tr
F
p
nd
/F
p
d
is a surjective F
p
dlinear mapping and
whose kernel is an F
p
dvector subspace of dimension n −1.
Proof. Clearly Tr
F
p
nd
/F
p
d
is additive. For t ∈ F
p
d we have t
p
d
= t, so F
p
dlinearity follows
from the formula
tu + (tu)
p
d
+ (tu)
p
2d
+· · · + (tu)
p
(n−1)d
= tu +tu
p
d
+tu
p
2d
+· · · +tu
p
(n−1)d
.
To see that Tr
F
p
nd
/F
p
d
is surjective, notice that Tr
F
p
nd
/F
p
d
(u) = 0 if and only if u is a root of
the polynomial
X +X
p
d
+X
p
2d
+· · · +X
p
(n−1)d
∈ F
p
d[X]
which has degree p
(n−1)d
and so has at most p
(n−1)d
< p
nd
roots in F
p
nd. This means that
ker Tr
F
p
nd
/F
p
d
cannot be the whole of F
p
nd. Tr
F
p
nd
/F
p
d
is surjective since its codomain has
dimension 1.
There is a multiplicative version of this construction. Consider the function
N: F
×
p
nd
−→F
×
p
nd
for which
N(u) = uu
p
d
u
p
2d
· · · u
p
(n−1)d
= uF
d
(u) F
2d
(u) · · · F
(n−1)d
(u).
Then we have
F
d
(N(u)) = u
p
d
u
p
2d
u
p
3d
· · · u
p
nd
= u
p
d
u
p
2d
u
p
3d
· · · u
p
(n−1)d
u
= uu
p
d
u
p
2d
u
p
3d
· · · u
p
(n−1)d
= N(u).
So by Proposition 5.24(i), N(u) ∈ F
p
d. By redeﬁning the codomain we obtain the relative norm
Norm
F
p
nd
/F
p
d
: F
×
p
nd
−→F
×
p
d
; Norm
F
p
nd
/F
p
d
(u) = uu
p
d
u
p
2d
· · · u
p
(n−1)d
.
5.27. Proposition. The relative norm Norm
F
p
nd
/F
p
d
is a surjective group homomorphism.
79
Proof. Multiplicativity is obvious. The kernel of Norm
F
p
nd
/F
p
d
consists of the roots in F
p
nd
of the polynomial
X
1+p
d
+···+p
(n−1)d
−1 ∈ F
p
d[X],
so
 ker Norm
F
p
nd
/F
p
d
 1 +p
d
+· · · +p
(n−1)d
=
p
nd
−1
p
d
−1
.
Hence
 imNorm
F
p
nd
/F
p
d
 =
p
nd
−1
 ker Norm
F
p
nd
/F
p
d

p
d
−1.
Since imNorm
F
p
nd
/F
p
d
F
×
p
d
, we also have
 imNorm
F
p
nd
/F
p
d
 p
d
−1,
therefore
imNorm
F
p
nd
/F
p
d
= F
×
p
d
.
Exercises on Chapter 5
5.1. Show that Proposition 5.13 also applies to an integral domain in place of a ﬁeld.
5.2. What happens to Theorem 5.20 if we try to take p = 2.
5.3. Let f(X) ∈ F
p
d[X] be an irreducible polynomial with deg f(X) = n. Find the splitting ﬁeld
of f(X). Deduce that for any other irreducible polynomial g(X) ∈ F
p
d[X] with deg g(X) = n,
the splitting ﬁelds of f(X) and g(X) over F
p
d agree.
5.4. Find the smallest Galois ﬁelds containing all the roots of the following polynomials, in
each case ﬁnd a primitive root of this Galois ﬁeld:
(a) X
8
−1 ∈ F
41
[X]; (b) X
8
−1 ∈ F
5
[X]; (c) X
8
−1 ∈ F
11
[X]; (d) X
8
−1 ∈ F
2
[X].
5.5. Let w ∈ F
×
p
d
be a primitive root. If ℓ < d, show that w / ∈ F
×
p
ℓ
. Deduce that deg
F
p
w = d
and d  φ(p
d
−1).
5.6. Let p > 0 be a prime. Suppose that d 1, and K/F
p
d is an extension. For a ∈ K, let
g
a
(X) = X
p
d
−X −a ∈ K[X].
(a) If the polynomial g
a
(X) is irreducible over K, show that the splitting ﬁeld E of g
a
(X)
over K is separable and Gal(E/K)
∼
= F
p
d. [Hint: show that if u ∈ E is a root of g
a
(X)
in an extension E/K, then so is u +t for every t ∈ F
p
.]
(b) If d = 1, show that g
a
(X) is irreducible over K if and only if it has no root in K.
(c) If K is a ﬁnite ﬁeld and d > 1, explain why g
a
(X) can never be irreducible over K.
5.7. Let p be an odd prime, d 1 and write q = p
d
.
(a) Consider {±1} = {1, −1} as a group under multiplication. Show that there is a unique
group homomorphism λ
q
: F
×
q
−→ {±1} which is characterized by the requirement
that for every u ∈ F
×
q
, λ
q
(u) = 1 if and only if u = v
2
for some v ∈ F
×
q
. Is λ
q
always
surjective?
80
(b) Consider the set of all squares in F
q
,
Σ
q
= {u
2
∈ F
q
: u ∈ F
q
} ⊆ F
q
.
Show that the number of elements of Σ
q
is Σ
q
 = (q +1)/2. Deduce that if t ∈ F
q
then
the set
t −Σ
q
= {t −u
2
∈ F
q
: u ∈ F
q
}
has t −Σ
q
 = (q + 1)/2 elements.
(c) If t ∈ F
q
, show that
Σ
q
∩ (t −Σ
q
) 1.
Deduce that every element of F
q
is either a square or can be written as the sum of two
squares.
(d) Deduce that the equation x
2
+y
2
+z
2
= 0 has at least one nontrivial solution in F
q
.
(e) What can you say about the case p = 2?
81
CHAPTER 6
A Galois Miscellany
In this chapter we will explore some miscellaneous topics in Galois Theory. Historically,
Galois Theory has always been an important tool in Number Theory and Algebra, stimulating
the development of subjects such as Group Theory, Ring Theory and such diverse areas as Dif
ferential Equations, Complex Analysis and Algebraic Geometry. Many of the ideas introduced
in this chapter are of great importance in these and other mathematical areas.
6.1. A proof of the Fundamental Theorem of Algebra
We will prove the Fundamental Theorem of Algebra for the complex numbers C. This proof
is essentially due to Gauss but he did not use the historically more recent Sylow theory. It is
interesting to compare the proof below with others which use the topology of the plane and circle
or Complex Analysis; our proof only uses the connectivity of the real line (via the Intermediate
Value Theorem) together with explicit calculations in C involving square roots.
6.1. Theorem (The Fundamental Theorem of Algebra). The ﬁeld of complex numbers C is
algebraically closed and R = C.
Proof. We know that [C : R] = 2, so C/R is algebraic. Let p(X) ∈ C[X] be irreducible.
Then any root u of p(X) in the algebraic closure C is algebraic over R, so in C[X] we have
p(X)  minpoly
R,u
(X). The splitting ﬁeld of p(X) over C is contained in the splitting ﬁeld E of
minpoly
R,u
(X)(X
2
+ 1) over R. Since C E, we have 2  [E : R] and so 2   Gal(E/R).
Now consider a 2Sylow subgroup P Gal(E/R) and recall that  Gal(E/R)/P is odd.
For the ﬁxed subﬁeld of P, we have
[E
P
: R] =
 Gal(E/R)
P
,
which shows that E
P
/R has odd degree. The Primitive Element Theorem 3.75 allows us to
write E
P
= R(v) for some v whose minimal polynomial over R must also have odd degree.
But by the Intermediate Value Theorem, every real polynomial of odd degree has a real root,
so irreducibility implies that v has degree 1 over R and therefore E
P
= R. This shows that
Gal(E/R) = P, hence Gal(E/R) is a 2group.
As C/R is a Galois extension, we can consider the normal subgroup Gal(E/C)▹Gal(E/R) for
which  Gal(E/R) = 2  Gal(E/C). We must show that  Gal(E/C) = 1, so suppose not. From
the theory of 2groups, there is a normal subgroup N ▹ Gal(E/C) of index 2, so we can consider
the Galois extension E
N
/C of degree 2. But from known properties of C (see Proposition 3.29),
every quadratic aX
2
+ bX + c ∈ C[X] has complex roots (because we can ﬁnd square roots
of every complex number). So we cannot have an irreducible quadratic polynomial in C[X].
Therefore  Gal(E/C) = 1 and E = C.
83
6.2. Cyclotomic extensions
We begin by discussing the situation for cyclotomic extensions over Q using material dis
cussed in Section 1.3. Let ζ
n
= e
2πi/n
, the standard primitive nth root of 1 in C. In Theo
rem 1.43, it was claimed that the irreducible polynomial over Q which has ζ
n
as a root was the
nth cyclotomic polynomial
Φ
n
(X) =
∏
t=1,...,n−1
gcd(t,n)=1
(X −ζ
t
n
).
6.2. Theorem. Let n 2. Then
• Q(ζ
n
) = Q[X]/(Φ
n
(X));
• [Q(ζ
n
) : Q] = φ(n);
• Gal(Q(ζ
n
)/Q)
∼
= (Z/n)
×
, where the element t
n
∈ (Z/n)
×
acts on Q(ζ
n
) by t
n
· ζ
n
= ζ
t
n
.
Proof. Since the complex roots of Φ
n
(X) are the powers ζ
t
n
with t = 1, . . . , n − 1 and
gcd(t, n) = 1, Q(ζ
n
) is the splitting ﬁeld of Φ
n
(X) over Q and indeed Q(ζ
n
) = Q(ζ
t
n
) whenever t
has the above properties and so ζ
t
n
is a primitive nth root of unity. The main step in the proof
is to show that Φ
n
(X) ∈ Z[X] is irreducible. To do this we will show that every power ζ
t
n
as
above is actually a Galois conjugate of ζ
n
over Q, therefore
Φ
n
(X) = minpoly
Q,ζ
n
(X) = minpoly
Q,ζ
t
n
(X)
and hence Φ
n
(X) is irreducible.
Consider
Z(ζ
n
) = {a
0
+a
1
ζ
n
+· · · +a
r
ζ
r
n
: r 0, a
j
∈ Z} ⊆ Q(ζ
n
).
Then Z(ζ
n
) is a subring of Q(ζ
n
) and so is an integral domain. Its group of units contains the
cyclic subgroup ⟨ζ
n
⟩ of order n.
Let p > 0 be a prime which does not divide n. Let P ▹ Z(ζ
n
) be a maximal ideal which
contains p; then the quotient ring Z(ζ
n
)/P is a ﬁeld of characteristic p. In fact, it is a ﬁnite
ﬁeld, say F
p
d for some d. Let π: Z(ζ
n
) −→F
p
d be the quotient homomorphism.
Inside the group of units of Z(ζ
n
) is the subgroup of powers of ζ
n
, ⟨ζ
n
⟩ Z(ζ
n
)
×
; this
is a cyclic subgroup of order n. We claim that when restricted to ⟨ζ
n
⟩, π gives an injective
group homomorphism, π
′
: ⟨ζ
n
⟩ −→ F
×
p
d
. To see this, suppose that π
′
(ζ
r
n
) = 1 for some r =
1, 2, . . . , n −1; then ζ
r
n
−1 ∈ P. By elementary Group Theory we can assume that r  n and so
p r. On factoring we have
(ζ
n
−1)(ζ
r−1
n
+· · · +ζ
n
+ 1) ≡ (ζ
n
−1)r (mod P),
so ζ
n
− 1 ∈ P or r ∈ P since maximal ideals are prime. But Z ∩ P = (p) and so r / ∈ P, hence
ζ
n
−1 ∈ P. Recalling that
ζ
n−1
n
+· · · +ζ
n
+ 1 = 0,
we see that n ∈ P and hence p  n, thus contradicting our original assumption on n. So π
′
is
injective.
Writing u = π
′
(u), we can consider the eﬀect of the absolute Frobenius map F: F
p
d −→F
p
d
on ζ
t
n
= ζ
t
n
,
F(ζ
t
n
) = (ζ
t
n
)
p
= ζ
tp
n
.
84
This shows that in the Galois extension F
p
d/F
p
, ζ
t
n
is conjugate to ζ
tp
n
; by iterating this we ﬁnd
that ζ
t
n
is conjugate to every power of the form ζ
tp
k
n
.
Now let t = 1, . . . , n −1 and gcd(t, n) = 1. Suppose there is a factorization
Φ
n
(X) = f(X) minpoly
Q,ζ
n
(X)
for some monic polynomial f(X) ∈ Z[X] and f(ζ
t
n
) = 0. Consider the prime power factorization
t = p
r
1
1
· · · p
r
m
m
, where the p
j
are primes with 2 p
1
< · · · < p
m
and r
j
1 with. Since
gcd(t, n) = 1 we also have p
j
n s.
Now consider a maximal ideal P
1
▹ Z[ζ
n
] containing p
1
. Reducing modulo P
1
and working
in the resulting extension F
p
d
1
1
/F
p
1
, we ﬁnd that ζ
n
is conjugate to ζ
p
r
1
1
n
. By separability and
the fact that the reduction map π
1
: Z[ζ
n
] −→F
p
d
1
1
is injective on the powers of ζ
n
, we ﬁnd that
f(ζ
p
r
1
1
n
) ̸= 0 and so f(ζ
p
r
1
1
n
) ̸= 0 in Z[ζ
n
]. This shows that minpoly
Q,ζ
n
(ζ
p
r
1
1
n
) = 0 and so ζ
p
r
1
1
n
is
conjugate to ζ
n
.
Repeating this argument starting with ζ
p
r
1
1
n
and using the prime p
2
we ﬁnd that
minpoly
Q,ζ
n
(ζ
p
r
1
1
p
r
2
2
n
) = 0
and so ζ
p
r
1
1
p
r
2
2
n
is conjugate to ζ
n
. Continuing in this fashion, for each j = 1, . . . , m we have
minpoly
Q,ζ
n
(ζ
p
r
1
1
p
r
2
2
···p
r
j
j
n
) = 0
and so ζ
p
r
1
1
···p
r
j
j
n
is conjugate to ζ
n
. When j = m, this shows that minpoly
Q,ζ
n
(ζ
t
n
) = 0. Hence
ζ
t
n
is conjugate to ζ
n
in the extension Q(ζ
n
)/Q.
6.3. Theorem. For n > 2, consider the cyclotomic extension Q(ζ
n
)/Q where ζ
n
= e
2πi/n
.
Then Q(ζ
n
)
R
̸= Q(ζ
n
). Furthermore,
Q(ζ
n
)
R
= Q(ζ
n
)
⟨( )⟩
= Q(ζ
n
+ζ
n
) = Q(cos(2π/n)),
and
[Q(cos(2π/n)) : Q] =
φ(n)
2
.
Proof. Recall that
Gal(Q(ζ
n
)/Q)
∼
= Z/n
×
,
where the residue class of r acts by sending ζ
n
to ζ
r
n
. Complex conjugation corresponds to the
residue class of −1 ≡ n −1 (mod n). Making use of the identities
e
θi
= cos θ + sin θ i, cos θ =
1
2
(e
θi
+e
−θi
),
we obtain
cos(2π/n) =
1
2
(ζ
n
+ζ
n
) =
1
2
(ζ
n
+ζ
−1
n
).
Complex conjugation ﬁxes each of the real numbers cos(2πk/n) for k = 1, 2, . . . , n − 1. The
residue class of r acts by sending cos(2π/n) to cos(2πr/n); it is elementary to show that
cos(2πr/n) ̸= cos(2π/n) unless r ≡ 1 (mod n). Hence
⟨
( )
⟩
= {id, ( )} = Gal(Q(cos(2π/n))/Q).
85
Thus we have
Q(ζ
n
)
⟨( )⟩
= Q(cos(2π/n)),
and so [Q(cos(2π/n)) : Q] = φ(n)/2. Notice that ζ
n
is a root of the polynomial
X
2
−2 cos(2π/n)X + 1 ∈ Q(cos(2π/n))[X],
so we also have
(6.1) minpoly
Q(cos(2π/n)),ζ
n
(X) = X
2
−2 cos(2π/n)X + 1.
6.4. Example. We have
[Q(ζ
24
) : Q] = φ(24) = 8
and
Gal(Q(ζ
24
)/Q)
∼
= Z/2 ×Z/2 ×Z/2.
Proof. By Theorem 1.43 we have [Q(ζ
24
) : Q] = 8. Also,
ζ
6
24
= i, ζ
3
24
=
√
2
2
+
√
2
2
i, ζ
8
24
= −
1
2
+
√
3
2
i,
and all of these numbers are in Q(ζ
24
), hence Q(
√
2,
√
3, i) Q(ζ
24
). It is easy to check that
[Q(
√
2,
√
3, i) : Q] = 8,
which implies that
Q(ζ
24
) = Q(
√
2,
√
3, i).
Using this we ﬁnd that
Gal(Q(ζ
24
)/Q)
∼
= Z/2 ×Z/2 ×Z/2.
We also have cos(2π/24) = cos(π/12) ∈ Q(ζ
24
). Since
cos(2π/12) = cos(π/6) =
√
3
2
,
we have
2 cos
2
(π/12) −1 =
√
3
2
and so
4 cos
4
(π/12) −4 cos
2
(π/12) + 1 =
3
4
,
giving
16 cos
4
(π/12) −16 cos
2
(π/12) + 1 = 0.
Then
16X
4
−16X
2
+ 1 = 16 minpoly
Q,cos(π/12)
(X).
Note that case (i) of Kaplansky’s Theorem 4.28 applies to the polynomial minpoly
Q,cos(π/12)
(X).
For this example, Gal(Q(ζ
24
)/Q) has 2
3
−1 = 7 subgroups of each of the orders 2 and 4; it
is an interesting exercise to ﬁnd them all together with their ﬁxed subﬁelds.
86
6.5. Remark. The minimal polynomial for cos(π/12) can also be found as follows. We have
Φ
24
(ζ
24
) = 0, hence since
Φ
24
(X) = X
8
−X
4
+ 1,
we obtain
ζ
8
24
−ζ
4
24
+ 1 = 0.
Then after multiplying by ζ
−4
24
we have
ζ
4
24
−1 +ζ
−4
24
= 0,
giving
(ζ
4
24
+ζ
−4
24
) −1 = 0.
Now
(ζ
24
+ζ
−1
24
)
4
= (ζ
4
24
+ζ
−4
24
) + 4(ζ
2
24
+ζ
−2
24
) + 6,
hence
ζ
4
24
+ζ
−4
24
= (ζ
24
+ζ
−1
24
)
4
−4(ζ
2
24
+ζ
−2
24
) −6.
Similarly,
(ζ
24
+ζ
−1
24
)
2
= ζ
2
24
+ζ
−2
24
+ 2,
so
ζ
2
24
+ζ
−2
24
= (ζ
24
+ζ
−1
24
)
2
−2.
Combining these we have
(ζ
24
+ζ
−1
24
)
4
−4(ζ
24
+ζ
−1
24
)
2
+ 1 = 0,
and so
16 cos
4
(π/12) −16 cos
2
(π/12) + 1 = 0.
This method will work for any n where φ(n) is even, i.e., when n > 2.
6.6. Remark. The polynomial that expresses cos nθ as a polynomial in cos θ is the n
th Chebsyhev polynomial of the ﬁrst kind T
n
(X) ∈ Z[X]. Here are the ﬁrst few of these
polynomials:
T
2
(X) = 2X
2
−1, T
3
(X) = 4X
3
−3X,
T
4
(X) = 8X
4
−8X
2
+ 1, T
5
(X) = 16X
5
−20X
3
+ 5X,
T
6
(X) = 32X
6
−48X
4
+ 18X
2
−1, T
7
(X) = 64X
7
−112X
5
+ 56X
3
−7X.
These form a system of orthogonal polynomials which can be computed in Maple using the
command orthopoly[T](n,X).
Now let K be a ﬁeld with characteristic char K n. The polynomial Φ
n
(X) has integer
coeﬃcients, so we can view it as an element of K[X] since either Q K or F
p
K and we
can reduce the coeﬃcients modulo p. In either case it can happen that Φ
n
(X) factors in K[X].
However, we can still describe the splitting ﬁeld of X
n
−1 over K and its Galois group.
6.7. Theorem. If char K n, then the splitting ﬁeld of X
n
−1 over K is K(ζ), where ζ ∈ K
is a primitive nth root of unity. The Galois group Gal(K(ζ)/K) is isomorphic to a subgroup
of (Z/n)
×
, hence it is abelian with order dividing φ(n).
87
Proof. Working in K, we know that Φ
n
(ζ) = 0, hence the roots of minpoly
K,ζ
(X) ∈ K[X]
are primitive roots of 1. So X
n
− 1 splits over K(ζ) and each element α ∈ Gal(K(ζ)/K) has
the action α(ζ) = ζ
r
α
, where gcd(r
α
, n) = 1. Hence Gal(K(ζ)/K) is isomorphic to a subgroup
of Gal(Q(ζ
n
)/Q)
∼
= (Z/n)
×
which implies that it is abelian and its order divides φ(n).
6.8. Remark. When p = char K > 0, this Galois group only depends on the largest subﬁeld
of K which is algebraic over F
p
. For example, if K = F
p
d(T) then the value of d is the crucial
factor. The precise outcome can be determined with the aid of Proposition 5.21.
6.9. Example. We have the following splitting ﬁelds and Galois groups.
(i) The splitting ﬁeld of X
4
−1 over F
3
(T) is F
9
(T) and
Gal(F
9
(T)/F
3
(T))
∼
= (Z/4)
×
∼
= Z/2.
(ii) By Proposition 5.20, X
4
−1 splits over F
5
(T) and the Galois group Gal(F
5
(T)/F
5
(T))
is trivial.
Proof. (i) By Proposition 5.20, X
4
−1 is separable over F
3
(T) and has irreducible factors
(X − 1), (X + 1) and (X
2
+ 1). The splitting ﬁeld of (X
2
+ 1) over F
3
is F
9
= F
3
(ζ), where
ζ
2
+ 1 = 0, so (X
2
+ 1) splits over F
9
(T). Also,
Gal(F
9
/F
3
)
∼
= (Z/4)
×
∼
= Z/2,
with generator σ satisfying σ(ζ) = ζ
−1
= −ζ. This generator clearly extends to an automor
phism of F
9
(T) which ﬁxes T.
(ii) By Proposition 5.20, X
4
−1 splits over F
5
.
6.3. Artin’s Theorem on linear independence of characters
Let G be a group and K a ﬁeld.
6.10. Definition. A group homomorphism χ: G −→ K
×
is called a character of G with
values in K.
6.11. Example. Given any ring homomorphism φ: R −→ K we obtain a character of R
×
in K by restricting φ to a map χ
φ
: R
×
−→ K
×
.
6.12. Example. Given an automorphism α: K −→ K, χ
α
: K
×
−→ K
×
is a character of
K
×
in K.
6.13. Example. Let E/K be a Galois extension and σ ∈ Gal(E/K). Then χ
σ
: E
×
−→ E
×
is a character.
6.14. Definition. Let χ
1
, . . . , χ
n
be characters of a group G in a ﬁeld K. Then χ
1
, . . . , χ
n
are linearly independent if for t
1
, . . . , t
n
∈ K,
t
1
χ
1
+· · · +t
n
χ
n
= 0 =⇒ t
1
= · · · = t
n
= 0.
If χ
1
, . . . , χ
n
are not linearly independent then they are linearly dependent.
In this deﬁnition, the functional equation means that for all g ∈ G,
t
1
χ
1
(g) +· · · +t
n
χ
n
(g) = 0.
88
6.15. Theorem (Artin’s Theorem). Let χ
1
, . . . , χ
n
be distinct characters of a group G in a
ﬁeld K. Then χ
1
, . . . , χ
n
are linearly independent.
Proof. We proceed by induction on n. For n = 1 the result is easily veriﬁed. For the
inductive assumption, suppose that it holds for any n k.
Let χ
1
, . . . , χ
k+1
be a set of k + 1 distinct characters for which there are t
1
, . . . , t
k+1
∈ K
not all zero and such that
(6.2) t
1
χ
1
+· · · +t
k+1
χ
k+1
= 0.
If one of the t
i
is zero, say t
r
= 0, then χ
1
, . . . , χ
r−1
, χ
r+1
, . . . , χ
k+1
is linearly dependent,
contradicting the inductive assumption. Hence all of the t
i
must be nonzero. As χ
1
̸= χ
2
,
there must be an element g
0
∈ G for which χ
1
(g
0
) ̸= χ
2
(g
0
). So for all g ∈ G, Equation (6.2)
applied to g
0
g yields
t
1
χ
1
(g
0
g) +· · · +t
k+1
χ
k+1
(g
0
g) = 0,
and therefore since χ
j
(g
0
g) = χ
j
(g
0
)χ
j
(g), we see that
t
1
χ
1
(g
0
)χ
1
+· · · +t
k+1
χ
k+1
(g
0
)χ
k+1
= 0.
Multiplying Equation (6.2) by χ
1
(g
0
) and subtracting gives
t
2
(χ
2
(g
0
) −χ
1
(g
0
))χ
2
+t
3
(χ
3
(g
0
) −χ
1
(g
0
))χ
3
+· · · +t
k+1
(χ
k+1
(g
0
) −χ
1
(g
0
))χ
k+1
= 0,
in which the coeﬃcient t
2
(χ
2
(g
0
)−χ
1
(g
0
)) is not zero. Hence χ
2
, . . . , χ
k+1
is linearly dependent,
again contradicting the inductive assumption. So χ
1
, . . . , χ
k+1
is linearly independent, which
demonstrates the inductive step.
6.16. Corollary. Suppose that α
1
, . . . , α
n
are distinct automorphisms of the ﬁeld K. Let
t
1
, . . . , t
n
∈ K be a sequence of elements, not all of which are 0. Then there is a z ∈ K for
which
t
1
α
1
(z) +· · · +t
n
α
n
(z) ̸= 0.
Hence the Klinear transformation t
1
α
1
+· · · +t
n
α
n
: K −→ K is nontrivial.
6.17. Corollary. Let E/K be a ﬁnite Galois extension of degree n and let α
1
, . . . , α
n
be the distinct elements of Gal(E/K). Then the function α
1
+ · · · + α
n
: E −→ E is a non
trivial Klinear transformation whose image is contained in K. Hence the associated Klinear
transformation
Tr
E/K
: E −→ K; Tr
E/K
(x) = α
1
(x) +· · · +α
n
(x)
is surjective.
The function Tr
E/K
: E −→ K is called the trace mapping of E/K.
Proof. First note that for x ∈ E and γ ∈ Gal(E/K),
γ(α
1
(x) +· · · +α
n
(x)) = γα
1
(x) +· · · +γα
n
(x) = α
1
(x) +· · · +α
n
(x),
since the list γα
1
, . . . , γα
n
is the same as α
1
, . . . , α
n
apart from its order. Hence,
α
1
(x) +· · · +α
n
(x) ∈ E
Gal(E/K)
= K.
The rest of the statement follows directly from Corollary 6.16.
89
Suppose that E/K is a ﬁnite Galois extension with cyclic Galois group Gal(E/K) = ⟨σ⟩ of
order n. For each u ∈ E
×
, the element uσ(u) · · · σ
n−1
(u) ∈ E satisﬁes
σ(uσ(u) · · · σ
n−1
(u)) = σ(u) · · · σ
n−1
(u)σ
n
(u) = σ(u) · · · σ
n−1
(u)u,
hence in uσ(u) · · · σ
n−1
(u) ∈ E
⟨σ⟩
= K. Now using this we deﬁne a group homomorphism
N
E/K
: E
×
−→ K
×
; N
E/K
(u) = uσ(u) · · · σ
n−1
(u).
N
E/K
is called the norm mapping for E/K and generalizes the norm mapping for ﬁnite ﬁelds
of Section 5.3.
There is another homomorphism
δ
E/K
: E
×
−→ E
×
; δ
E/K
(u) = uσ(u)
−1
.
Notice that for u ∈ E
×
,
N
E/K
(δ
E/K
(u)) = (uσ(u)
−1
)(σ(u)σ
2
(u)
−1
· · · σ
n−1
(u)σ
n
(u)
−1
) = 1,
since σ
n
(u) = u. So imδ
E/K
ker N
E/K
. Our next result is an important generalization of
Proposition 5.27.
6.18. Theorem (Hilbert’s Theorem 90). Let E/K be a ﬁnite Galois extension with cyclic
Galois group Gal(E/K) = ⟨σ⟩ of order n. Then imδ
E/K
= ker N
E/K
. Explicitly, if u ∈ E
×
and
uσ(u) · · · σ
n−1
(u) = 1, then there is a v ∈ E
×
such that u = vσ(v)
−1
.
Proof. Let u ∈ ker N
E/K
.
The characters σ
k
: E
×
−→ E
×
with k = 0, 1, . . . , n−1 are distinct and linearly independent
by Artin’s Theorem 6.15. Consider the function
id +uσ +uσ(u)σ
2
+· · · +uσ(u) · · · σ
n−2
(u)σ
n−1
: E
×
−→ E.
This cannot be identically zero, so for some w ∈ E, the element
v = w +uσ(w) +uσ(u)σ
2
(w) +· · · +uσ(u) · · · σ
n−2
(u)σ
n−1
(w)
is nonzero. Notice that
uσ(v) = uσ(w) +uσ(u)σ
2
(w) +uσ(u)σ
2
(u)σ
3
(w) +· · · +uσ(u)σ
2
(u) · · · σ
n−1
(u)σ
n
(w) = v,
since
uσ(u)σ
2
(u) · · · σ
n−1
(u)σ
n
(w) = w.
Thus we have u = vσ(v)
−1
as required.
6.4. Simple radical extensions
In this section we will investigate splitting ﬁelds of polynomials of the form X
n
−a, where
char K n. We call these simple radical extensions and later in Deﬁnition 6.33 we introduce a
more general notion of radical extension.
6.19. Proposition. Let f(X) = X
n
−a ∈ K[X] be irreducible and separable over K. Then
the splitting ﬁeld of f(X) over K has the form K(u, ζ), where u is a root of f(X) and ζ is a
primitive nth root of 1.
90
6.20. Corollary. If K contains a primitive nth root of 1, ζ, then the splitting ﬁeld of
f(X) = X
n
− a over K has the form K(u), where u is a root of f(X). The Galois group
Gal(K(u)/K) is cyclic of order n with a generator σ for which σ(u) = ζu.
In the more general situation of Proposition 6.19,
{id} ▹ Gal(K(ζ, u)/K(ζ)) ▹ Gal(K(ζ, u)/K),
where Gal(K(ζ, u)/K(ζ)) is cyclic and
Gal(K(ζ)/K))
∼
= Gal(K(ζ, u)/K))/ Gal(K(ζ, u)/K(ζ))
is abelian. The Galois Correspondence identiﬁes the following towers of subﬁelds and subgroups.
K(ζ, u)
Gal(K(ζ, u)/K)
▹
.
K(ζ)
Gal(K(ζ, u)/K(ζ))
K {id}
6.21. Definition. Let K be a ﬁeld with char K n and which contains a primitive nth
root of 1, ζ say. Then L/K is a simple nKummer extension if L = K(u) where u
n
= a
for some a ∈ K. L/K is an (iterated) nKummer extension if L = K(u
1
, . . . , u
k
) where
u
n
1
= a
1
, . . . , u
n
k
= a
k
for some elements a
1
, . . . , a
k
∈ K.
Note that in this deﬁnition we do not require the polynomials X
n
− a
j
∈ K[X] to be
irreducible.
6.22. Proposition. Let K(u)/K be a simple nKummer extension. Then K(u)/K is a
Galois extension and Gal(K(u)/K) is cyclic with order dividing n.
Proof. Suppose that u
n
= a ∈ K. Then in K[X] we have
X
n
−a = (X −u)(X −ζu) · · · (X −ζ
n−1
u).
Clearly the roots of X
n
−a are distinct and so K(u)/K is separable over K; in fact, K(u) is a
splitting ﬁeld of X
n
−a over K. This means that K(u)/K is Galois.
For each α ∈ Gal(K(u)/K) we have α(u) = ζ
r
α
u for some r
α
= 0, 1 . . . , n −1. Notice that
for β ∈ Gal(K(u)/K),
βα(u) = β(ζ
r
α
u) = ζ
r
α
β(u) = ζ
r
α
ζ
r
β
u = ζ
r
α
+r
β
u,
and so r
βα
= r
α
+r
β
. Hence the function
ρ: Gal(K(u)/K) −→ ⟨ζ⟩ ; ρ(α) = ζ
r
α
,
is a group homomorphism. As ⟨ζ⟩ is cyclic of order n, Lagrange’s Theorem implies that the
image of ρ has order dividing n. Since every element of Gal(K(u)/K) is determined by its eﬀect
on u, ρ is injective, hence  Gal(K(u)/K) divides n. In fact, Gal(K(u)/K) is cyclic since every
subgroup of a cyclic group is cyclic.
6.23. Example. Let n 1 and q ∈ Q. Then Q(ζ
n
,
n
√
q)/Q(ζ
n
) is a simple nKummer
extension.
91
6.24. Example. Q(i,
√
2)/Q(i) is a simple 4Kummer extension with Gal(Q(i,
√
2)/Q(i))
cyclic of order 2.
Proof. We have (
√
2)
4
−4 = 0, but
X
4
−4 = (X
2
−2)(X
2
+ 2),
and
X
2
−2 = minpoly
Q(i),
√
2
(X).
The corresponding group homomorphism ρ: Gal(Q(i)(
√
2)/Q(i)) −→ ⟨i⟩ has image
imρ = {1, −1} ⟨i⟩ .
Here is a converse to Proposition 6.22.
6.25. Proposition. Suppose that char K n and there is an element ζ ∈ K which is a
primitive nth root of unity. If E/K is a ﬁnite Galois extension with cyclic Galois group of
order n, then there is an element a ∈ E such that E = K(a) and a is a root of a polynomial of
the form X
n
−b with b ∈ K. Hence E/K is a simple nKummer extension.
Proof. We have
N
E/K
(ζ
−1
) = ζ
−n
= 1,
so by Hilbert’s Theorem 6.18, there is an element a ∈ E for which ζ
−1
= aσ(a)
−1
. Then
σ(a) = ζa and the elements σ
k
(a) = ζ
k
a for k = 0, 1, . . . , n − 1 are distinct, so they must be
the n conjugates of a. Also note that
X
n
−a
n
= (X −a)(X −ζa) · · · (X −ζ
n−1
a) = (X −a)(X −σ(a)) · · · (X −σ
n−1
(a)),
hence a
n
∈ K since it is ﬁxed by σ. Since K(a) E, this shows that
n = [K(a) : K] [E : K] = n
and therefore
[K(a) : K] = [E : K] = n,
whence K(a) = E.
6.5. Solvability and radical extensions
We begin by recalling some ideas about groups, see [3, 5] for further details.
6.26. Definition. A group G is solvable, soluble or soluable if there is a chain of subgroups
(called a subnormal series)
{1} = G
ℓ
G
ℓ−1
· · · G
1
G
0
= G
in which G
k+1
▹ G
k
and each composition factor G
k
/G
k+1
is abelian; we usually write
{1} = G
ℓ
▹ G
ℓ−1
▹ · · · ▹ G
1
▹ G
0
= G.
If each composition factor is a cyclic group of prime order the subnormal series is called a
composition series. A group which is not solvable is called insolvable.
92
6.27. Remark. It is a standard result that we can always reﬁne (i.e., add extra terms) a
subnormal series of a solvable group to obtain a composition series. The primes appearing as
well as the number of times each occurs are all determined by G, only their order varying for
diﬀerent composition series.
6.28. Example. Let G be a ﬁnite abelian group. Then G is solvable.
6.29. Example. Let G be a ﬁnite pgroup, where p is a prime. Then G is solvable.
In fact, for a ﬁnite pgroup G, there is always a normal subgroup of a pgroup with index
p, so in this case we can assume each quotient G
k
/G
k+1
is cyclic of order p.
6.30. Proposition. Let G be a group.
(i) If G is solvable then every subgroup H G and every quotient group G/N is solvable.
(ii) If N ▹ G and G/N are solvable then so is G.
In the opposite direction we can sometimes see that a group is insolvable. Recall that a
group is simple if it has no nontrivial proper normal subgroups.
6.31. Proposition. Let G be a ﬁnite group. Then G is insolvable if any of the following
conditions holds:
(i) G contains a subgroup which is a nonabelian simple group.
(ii) G has a quotient group which is a nonabelian simple group.
(iii) G has a composition series in which one of the terms is a nonabelian simple group.
6.32. Example. For n 5, the alternating and symmetric groups A
n
and S
n
are insolvable.
Proof. This follows from the fact that if n 5, A
n
is a simple group and A
n
▹ S
n
with
quotient group S
n
/A
n
∼
= Z/2.
Now we explain how this relates to ﬁelds and their extensions. Let K be a ﬁeld and L/K a
ﬁnite extension. For simplicity, we assume also that char K = 0.
6.33. Definition. L/K is a radical extension of K if it has the form L = K(a
1
, a
2
, . . . , a
n
)
with
a
d
k
k
∈ K(a
1
, a
2
, . . . , a
k−1
)
for some d
k
1. Thus every element of L is expressible in terms of iterated roots of elements
of K.
We will need the following Lemma and its Corollary. According to [4], several text books
make subtle errors or omissions related to this result, so beware when reading other sources!
6.34. Lemma. Let L/K be a ﬁnite Galois extensions and let L(u)/L be a radical extension.
Let E/L be an extension where E is a splitting ﬁeld for the polynomial minpoly
K,u
(X) ∈ L[X].
Then E/L is a radical Galois extension. In particular, if L/K is a radical Galois extension
then so is E/K.
Proof. Suppose that u
d
= a ∈ L with a ̸= 0. Then X
d
− a has a d distinct roots in E,
and if v is any other root then (v/u)
d
= 1, so there are d distinct dth roots of unity in E.
Hence there is a primitive dth root of unity ζ ∈ E and the subﬁeld L(ζ, u) E is normal over
93
L, so L(ζ, u)/L is a radical Galois extension. But L(ζ, u)/K need not be Galois. However, if
u = u
1
, . . . , u
t
∈ E are the distinct roots of minpoly
K,u
(X) in E, then
E = L(ζ, u, u
1
, . . . , u
t
).
But this is clearly a radical extension of L.
If L/K is a radical Galois extension, say L = K(a
1
, . . . , a
n
), then
E = L(a
1
, . . . , a
n
, ζ, u, u
1
, . . . , u
t
),
which is a radical Galois extension of K.
6.35. Corollary. If L/K is a radical extension then it is contained in a radical Galois
extension L
′
/K.
Proof. Writing L = K(a
1
, a
2
, . . . , a
n
) as in Deﬁnition 6.33, this is proved by induction
on n using Lemma 6.34.
In the next deﬁnition, the word Galois is superﬂuous because of the preceding results.
6.36. Definition. If L is the splitting ﬁeld of a polynomial f(X) ∈ K[X], then f(X) is
solvable by radicals over K if L is contained in a radical (Galois) extension of K.
6.37. Definition. L/K is solvable if L L
′
where L
′
/K is a ﬁnite radical Galois extension
of K.
6.38. Theorem. Let E/K be a ﬁnite Galois extension. Then E/K is solvable if and only
if the group Gal(E/K) is solvable.
Proof. Suppose that E E
′
where E
′
/K is a ﬁnite radical Galois extension, so
E
′
= K(ζ, u
1
, . . . , u
m
),
where ζ
d
= 1, u
d
1
1
∈ K(ζ) and u
d
r
r
∈ K(ζ, u
1
, . . . , u
r−1
) for r = 2, . . . , m with d
1
· · · d
m
 d. If
G
r
▹Gal(E
′
/K) and
(E
′
)
G
r
= K(ζ, u
1
, . . . , u
r
),
with
(E
′
)
G
0
= K(ζ),
then
{1} = G
m
▹G
m−1
▹· · · ▹G
0
▹Gal(E
′
/K)
and
G
r−1
/G
r
∼
= Gal(K(ζ, u
1
, . . . , u
r
)/K(ζ, u
1
, . . . , u
r−1
)),
which is abelian by Proposition 6.22. Hence Gal(E
′
/K) is solvable, and since Gal(E/K) is a
quotient group of Gal(E
′
/K), is also solvable by Proposition 6.30.
Now suppose that Gal(E/K) is solvable and let n =  Gal(E/K). Let E
′
be the splitting
ﬁeld of X
n
−1 over E, so E
′
contains a primitive nth root of unity ζ and therefore it contains
a primitive dth root of unity for every divisor d of n. Now Gal(E
′
/E) ▹ Gal(E
′
/K) and
by Theorem 6.7, Gal(E
′
/E) is abelian. Also, Gal(E
′
/K)/ Gal(E
′
/E)
∼
= Gal(E/K) which is
solvable, so Gal(E
′
/K) is solvable by Proposition 6.30. We will now show that E
′
/K is a
radical extension.
94
Clearly K(ζ)/K is radical. Then Gal(E
′
/K(ζ)) ▹ Gal(E
′
/K) is solvable. Let
{1} = G
ℓ
▹ G
ℓ−1
▹ · · · ▹ G
1
▹ G
0
= Gal(E
′
/K(ζ))
be a composition series. The extension (E
′
)
G
1
/K(ζ) is radical by Proposition 6.25. Similarly,
each extension (E
′
)
G
k+1
/(E
′
)
G
k
is radical. Hence E
′
/K(ζ) is radical, as is E
′
/K.
6.39. Example. The Galois group of the extension Q(ζ
3
,
3
√
2)/Q is solvable.
Proof. We have already studied this extension in Example 3.30 and 4.20. Clearly Q(ζ
3
,
3
√
2)
is a radical extension of Q and
Q(ζ
3
,
3
√
2) = Q(ζ
3
)(
3
√
2).
We know that Gal(Q(ζ
3
,
3
√
2)/Q)
∼
= S
3
, where we identify each element of the Galois group with
a permutation of the three roots of X
3
−2 in Q(ζ
3
,
3
√
2) which we list in the order
3
√
2,
3
√
2 ζ
3
,
3
√
2 ζ
2
3
.
We have the following towers of subﬁelds and subgroups related under the Galois Correspon
dence.
Q(ζ
3
,
3
√
2)
_
S
3
Q(ζ
3
) = Q(ζ
3
,
3
√
2)
A
3
3
A
3
= Gal(Q(ζ
3
,
3
√
2)/Q(ζ
3
))
2
Q
2
{id}
3
Here Q(ζ
3
)/Q is itself a Galois extension and A
3
▹ S
3
. Notice that A
3
∼
= Z/3 and S
3
/A
3
∼
= Z/2,
so we have the following composition series for S
3
:
{id} ▹ A
3
▹ S
3
.
It is also interesting to reverse the question and ask whether there are extensions which are
not solvable. This was a famous problem pursued for several hundred years. To ﬁnd examples,
we ﬁrst recall that the smallest nonabelian simple group is A
5
which has order 60. We should
therefore expect to look for a polynomial of degree at least 5 to ﬁnd a Galois group for a splitting
ﬁeld to be simple or occur as a composition factor of such a Galois group. Here is an explicit
example over Q.
6.40. Example. The splitting ﬁeld of the polynomial f(X) = X
5
−35X
4
+7 ∈ Q[X] is not
solvable.
Proof. Let E C be the splitting ﬁeld of f(X) over Q. Using the Eisenstein Test 1.38
with p = 7, we ﬁnd that f(X) is irreducible over Q. By Theorem 4.8(iii), 5 divides the order of
Gal(E/Q), so by Cauchy’s Lemma this group contains an element of order 5.
Now observe that
f
′
(X) = 5X
4
−140X
3
= 5X
3
(X −28), f
′′
(X) = 20X
4
−420X
2
= 20X
2
(X −21).
There are two turning points, namely a maximum at x = 0 and a minimum at x = 28. Then
f(0) = 7 > 0 > f(28) = −4302585,
95
hence there are three real roots of f(X) and two nonreal complex ones. Then complex conju
gation restricts to an element of order 2 in Gal(E/Q) which interchanges the nonreal roots and
ﬁxes the others. If we list the roots of f(X) as u
1
, u
2
, u
3
, u
4
, u
5
with u
1
, u
2
being the nonreal
roots, then the transposition (1 2) ∈ S
5
corresponds to this element. Furthermore, the only
elements of S
5
of order 5 are 5cycles; by taking an appropriate power we can assume that there
is a 5cycle of the form (1 2 3 4 5) corresponding to an element of Gal(E/Q) which we can view
as a subgroup of S
5
. The next lemma shows that Gal(E/Q)
∼
= S
5
.
6.41. Lemma. Let n 1. Suppose that H S
n
and H contains the elements (1 2) and
(1 2 · · · n). Then H = S
n
.
The proof is left as an exercise. This completes the veriﬁcation of Example 6.40.
It is worth remarking that the most extreme version of this occurs when we ask for a Galois
group which is simple. There has been a great deal of research activity on this question in the
past few decades, but apparently not all simple groups are known to occur as Galois groups of
extensions of Q or other ﬁnite subextensions of C/Q. Here is an example whose Galois group
is A
5
; this is veriﬁed using Proposition 4.26.
6.42. Example. The Galois group of f(X) = X
5
+20X +16 over Q is Gal(Q(f(X))/Q)
∼
=
A
5
, hence it is not solvable.
6.6. Symmetric functions
Let k be a ﬁeld. Consider the polynomial ring on n indeterminates k[X
1
, . . . , X
n
] and its
ﬁeld of fractions K = k(X
1
, . . . , X
n
). Each permutation σ ∈ S
n
acts on k[X
1
, . . . , X
n
] by
σ · f(X
1
, . . . , X
n
) = f
σ
(X
1
, . . . , X
n
) = f(X
σ(1)
, . . . , X
σ(n)
).
Viewed as a function σ· : k[X
1
, . . . , X
n
] −→k[X
1
, . . . , X
n
] is a ring isomorphism; this extends to
a ring isomorphism σ· : k(X
1
, . . . , X
n
) −→ k(X
1
, . . . , X
n
). Varying σ we obtain actions of the
group S
n
on k[X
1
, . . . , X
n
] and k(X
1
, . . . , X
n
) by ring isomorphisms ﬁxing k and in the latter
case it is by ﬁeld automorphisms ﬁxing k.
6.43. Definition. The ﬁeld of symmetric functions on n indeterminates is
Sym
n
(k) = k(X
1
, . . . , X
n
)
S
n
k(X
1
, . . . , X
n
).
So if f(X
1
, . . . , X
n
) ∈ k(X
1
, . . . , X
n
), then
f(X
1
, . . . , X
n
) ∈ Sym
n
(k) ⇐⇒ ∀σ ∈ S
n
f(X
1
, . . . , X
n
) = f(X
σ(1)
, . . . , X
σ(n)
).
6.44. Theorem. The extension k(X
1
, . . . , X
n
)/ Sym
n
(k) is a ﬁnite Galois extension for
which Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k))
∼
= S
n
.
Proof. There are elements of k[X
1
, . . . , X
n
] ⊆ k(X
1
, . . . , X
n
) called elementary symmetric
functions,
e
k
=
∑
i
1
<i
2
<···<i
k
X
i
1
X
i
2
· · · X
i
k
,
where 1 k n. It is easy to see that for every σ ∈ S
n
, e
σ
k
= e
k
, so e
k
∈ Sym
n
(k). Working in
the ring k(X
1
, . . . , X
n
)[Y ] we have
f
n
(Y ) = Y
n
−e
1
Y
n−1
+· · · + (−1)
n−1
e
n−1
Y + (−1)
n
e
n
= 0,
96
hence the roots of this polynomial are the X
i
. So k(X
1
, . . . , X
n
) is the splitting ﬁeld of f
n
(Y )
over Sym
n
(k). Now S
n
Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k)), hence
[k(X
1
, . . . , X
n
) : Sym
n
(k)] =  Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k)) S
n
 = n!.
But as every element of Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k)) permutes the roots of f
n
(Y ) and is de
termined by this permutation, we also have
n!  Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k)).
Combining these inequalities we obtain  Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k)) = n! and therefore
Gal(k(X
1
, . . . , X
n
)/ Sym
n
(k)) = S
n
.
6.45. Remark. In fact, this proof shows that the extension k(X
1
, . . . , X
n
)/k(e
1
, . . . , e
n
) is
Galois of degree n!. Since k(e
1
, . . . , e
n
) Sym
n
(k) we can also deduce that k(e
1
, . . . , e
n
) =
Sym
n
(k). Hence every element of Sym
n
(k) is a rational function in the e
i
. Analogous results
are true for polynomials, i.e.,
k[X
1
, . . . , X
n
]
S
n
= k[e
1
, . . . , e
n
].
6.46. Corollary. If n 5, the extension k(X
1
, . . . , X
n
)/ Sym
n
(k) is not solvable.
Exercises on Chapter 6
6.1. Let p > 0 be a prime and G a group of order G = p
n
for some n 1. Show by induction
on n that there is a normal subgroup N ▹ G with N = p
n−1
. [Hint: what do you know about
the centre of G? Use this information to produce a quotient group of smaller order than G.]
6.2. Let K be a ﬁeld for which char K ̸= 2 and n 1 be odd. If K contains a primitive nth
root of unity, show that then K contains a primitive 2nth root of unity.
6.3. Find all values of n 1 for which φ(n)  4. Using this, determine which roots of unity lie
in the following ﬁelds:
Q(i), Q(
√
2 i), Q(
√
3 i), Q(
√
5 i).
6.4. (a) Describe the elements of (Z/24)
×
explicitly and verify that this group is isomorphic
to Z/2 ×Z/2 ×Z/2. Describe the eﬀect of each element on Q(ζ
24
) and Q(cos(π/12)) under the
action described in Theorem 6.2.
(b) Determine the group (Z/20)
×
and describe the eﬀect of each of its elements on Q(ζ
20
) and
Q(cos(π/10)) under the action described in Theorem 6.2.
6.5. Let n 1.
(a) What can you say about sin(2π/n) and Gal(Q(sin(2π/n))/Q))?
(b) Determine sin(π/12) and Gal(Q(sin(π/12))/Q)).
6.6. In this question, work in the cyclotomic ﬁeld Q(ζ
5
) where ζ
5
= e
2πi/5
.
(a) Describe the Galois group Gal(Q(ζ
5
)/Q) and its action on Q(ζ
5
).
97
(b) Determine the minimal polynomial of cos(2π/5) over Q. Hence show that
cos(2π/5) =
−1 +
√
5
4
.
For which other angles θ is cos θ a root of this minimal polynomial? What is the value
of sin(2π/5) ?
(c) Find the tower of subﬁelds of Q(ζ
5
) and express them as ﬁxed ﬁelds of subgroups of
Gal(Q(ζ
5
)/Q).
6.7. In this question, let p be an odd prime and let ζ
p
= e
2πi/p
∈ Q(ζ
p
) C.
(a) Consider the product
ξ =
(p−1)/2
∏
r=1
(ζ
r
p
−ζ
−r
p
) ∈ Q(ζ
p
).
Show that
ξ
2
= (−1)
(p−1)/2
p−1
∏
r=1
(1 −ζ
r
p
).
(b) Deduce that
ξ
2
=
p if p ≡ 1 (mod 4),
−p if p ≡ 3 (mod 4).
(c) Conclude that
ξ =
±
√
p if p ≡ 1 (mod 4),
±
√
p i if p ≡ 3 (mod 4).
and also
√
p ∈ Q(ζ
p
) if p ≡ 1 (mod 4) and
√
p i ∈ Q(ζ
p
) if p ≡ 3 (mod 4).
6.8. Prove Lemma 6.41. [Hint: show that every 2cycle of the form (i i + 1) is in H by
considering elements of the form (1 2 · · · n)
r
(1 2)(1 2 · · · n)
n−r
.]
6.9. This question is about an additive version of Hilbert’s Theorem 90, see Theorem 6.18.
Let E/K be a Galois extension with cyclic Galois group Gal(E/K) = ⟨σ⟩ of order n.
(a) Show that the function
T : E −→ E; T(u) = u +σ(u) +σ
2
(u) +· · · +σ
n−1
(u),
takes values in K and use this to deﬁne a Klinear mapping Tr
E/K
: E −→ K.
(b) If v ∈ E has Tr
E/K
(v) = 0, show that there is a w ∈ E such that v = w −σ(w).
[Hint: Show that there is an element t ∈ E for which Tr
E/K
t ̸= 0, then consider
w =
1
(Tr
E/K
t)
vσ(t) + (v +σ(v))σ
2
(t) +· · · + (v +σ(v)σ
2
(t) +· · · +σ
n−2
(v))σ
n−1
(t)
and adapt the proof of Hilbert’s Theorem 90 in Theorem 6.18, using Tr
E/K
in place of N
E/K
.]
6.10. (a) For n 1 and 1 k n, the kth power sum s
k
∈ k[X
1
, . . . , X
n
]
S
n
is deﬁned by
s
k
=
∑
1in
X
k
i
.
Prove the formula
s
k
= e
1
s
k−1
−e
2
s
k−2
+· · · + (−1)
k−1
e
k−1
s
1
+ (−1)
k
ke
k
.
98
(b) For n 1 and 1 k n, the total symmetric function is deﬁned by
h
k
=
∑
j
1
j
2
···j
k
X
j
1
X
j
2
· · · X
j
k
,
i.e., the sum of all the monomials in the X
i
of degree k.
(i) For large values of n, express h
1
, h
2
, h
3
in terms of the elementary symmetric functions
e
1
, e
2
, e
3
.
(ii) Show that the power sum functions s
k
of the previous question satisfy
s
k
= −(h
1
s
k−1
+h
2
s
k−2
+· · · +h
k−1
s
1
) +kh
k
.
99
Bibliography
[1] E. Artin, Galois Theory, Dover Publications (1998); ISBN 0 486 62342 4.
[2] JP. Escoﬁer, Galois theory, SpringerVerlag, New York (2001); ISBN 0387987657. [Highly recommended,
especially for its historical notes]
[3] J. B. Fraleigh, A First Course in Abstract Algebra, Addison Wesley (1999); ISBN 0 201 33596 4. [Highly
recommended]
[4] T. W. Hungerford, A counterexample in Galois theory, American Mathematical Monthly 97 (1997), 54–57.
[5] S. Lang, Algebra, Addison Wesley (1993); ISBN 0 201 55540 9.
[6] R. Lidl & H. Niederreiter, Finite Fields, Cambridge University Press (1997); ISBN 0 521 39231 4.
[7] J. Rotman, Galois Theory, SpringerVerlag (1998); ISBN 0 387 98541 7.
[8] I. Stewart, Galois Theory, Chapman and Hall (1989); ISBN 0 412 345501. [Very highly recommended.]
101
Solutions
Chapter 1
1.1. Clearly {n ∈ Z : n > 0 and nr = 0 for all r ∈ R} ⊆ {n ∈ Z : n > 0 and n1 = 0}. If
0 < n ∈ Z and n1 = 0, then for every r ∈ R,
nr = r +· · · +r
. .. .
n
= (1 +· · · + 1
. .. .
n
)r = (n1)r = 0r = 0,
so {n ∈ Z : n > 0 and n1 = 0} ⊆ {n ∈ Z : n > 0 and nr = 0 for all r ∈ R}. Hence these sets
are in fact equal. When char R = p > 0 they must both be nonempty. Now by deﬁnition of
characteristic,
char R = min{n ∈ Z : n > 0 and n1 = 0} = min{n ∈ Z : n > 0 and nr = 0 for all r ∈ R}.
1.2. (a) Let u, v ∈ S and suppose that uv = 0; then u = 0 or v = 0 since u, v ∈ R and R is
an integral domain. Consider the unit homomorphisms η: Z −→ R and η
′
: Z −→ S. Then for
n ∈ Z, η
′
(n) = η(n), so ker η
′
= ker η and therefore char S = char R.
(b) Q is a ﬁeld and Z ⊆ Q is a subring which is not a ﬁeld.
1.3. (a) For any subring R ⊆ C, R is an integral domain with characteristic subring Z and
char R = 0.
(b) The characteristic subring of A[X] is the same as that of A and char A[X] = char A. A[X]
is an integral domain if and only if A is an integral domain.
(c) If we identify A with the subring of scalar matrices in Mat
n
(A), then the characteristic
subring of Mat
n
(A) is the same as that of A and char Mat
n
(A) = char A. If n > 1 then
Mat
n
(A) is not commutative, in any case it always has zerodivisors since any singular matrix
is a zerodivisor.
1.4. The main thing to check is that φ(u + v) = φ(u) + φ(v) which is a consequence of the
Idiot’s Binomial Theorem. For R = F
p
[X], φ is not surjective, while for R = F
p
[X]/(X
2
), φ is
not injective.
1.5. (a) Recall from the Isomorphism Theorems of basic Ring Theory that φ
−1
Q▹ R; we need
to show it is a prime ideal. Suppose that u, v ∈ R with uv ∈ φ
−1
Q; then φ(u)φ(v) = φ(uv) ∈ Q
and so φ(u) ∈ Q or φ(v) ∈ Q, hence u ∈ φ
−1
Q or v ∈ φ
−1
Q.
(b) Consider the inclusion function inc: R −→ S; then inc
−1
Q = Q∩ R, so this result follows
from (a).
(c) Consider Z ⊆ Q; then the zeroideal (0)
Q
▹Q has (0)
Q
∩Z = (0)
Z
▹Z but this is not maximal
in Z since for any prime p > 0, (p)
Z
▹ Z is a (maximal) ideal that properly contains (0)
Z
.
(d) We have P ⊆ Q∩ R ▹ R with P ▹ R maximal; so P ⊆ Q∩ R. In fact Q only needs to be a
proper ideal of S for this argument to work.
1.6. The only proper ideal of k is the zero ideal (0), so ker φ = (0).
1.7. (a) Addition and multiplication follow from the obvious formulae
(u
1
+v
1
i)+(u
2
+v
2
i) = (u
1
+u
2
)+(v
1
+v
2
)i, (u
1
+v
1
i)(u
2
+v
2
i) = (u
1
u
2
−v
1
v
2
)+(u
1
v
2
+u
2
v
1
)i,
with Z[i] and Q[i] both closed under these operations and containing 1 = 1 + 0i as a unity, so
they are subrings of the ﬁeld C; by Qu. 1.1, they are both integral domains. To see that Q[i] is
a ﬁeld, notice that if u +vi ̸= 0 with u, v ∈ Q,
(u −vi)(u +vi) = (u +vi)(u −vi) = u
2
+v
2
̸= 0,
so
u
u
2
+v
2
+
v
u
2
+v
2
i ∈ Q(i)
is the inverse of u +vi. Hence every nonzero element of Q[i] has an inverse, therefore Q[i] is a
ﬁeld.
(b) & (c) The crucial point is that every element of Q[i] can be written as
1
n
(u + vi) with
n, u, v ∈ Z and n ̸= 0. Then
inc
∗
(u +vi)
n
= inc
∗
(u +vi)
n + 0i
=
1
n
(u +vi),
so the latter element is in the image of inc
∗
which must therefore be a surjection.
1.8. (a) Existence and uniqueness of such an ψ
a,b
follow from the Homomorphism Extension
Property 1.22 and its eﬀect on f(X) =
∑
n
i=0
r
i
X
i
∈ R[X] where r
i
∈ R is
ψ
a,b
(f(X)) = f(aX +b) =
n
∑
i=0
r
i
(aX +b)
i
.
We have
ψ
a,b
◦ ψ
c,d
(X) = ψ
a,b
(cX +d) = c(aX +b) +d = caX + (cb +d) = ψ
ca,cb+d
(X).
By the uniqueness part of the Homomorphism Extension Property, we have ψ
a,b
◦ψ
c,d
= ψ
ca,cb+d
.
If a is a unit then ψ
a
−1
,−ba
−1 : R[X] −→ R[X] has the property that ψ
a
−1
,−ba
−1(aX+b) = X and
ψ
a,b
(a
−1
X −ba
−1
) = X, so by the uniqueness part of the Homomorphism Extension Property,
ψ
a,b
◦ ψ
a
−1
,−ba
−1 = id = ψ
a
−1
,−ba
−1 ◦ ψ
a,b
.
Therefore these are inverse isomorphisms, ψ
a
−1
,−ba
−1 = ψ
−1
a,b
.
(b) (i) If f(X) =
∑
n
i=0
c
i
X
i
∈ k[X] with c
i
∈ k and c
n
̸= 0, then deg f(X) = n. Now
ψ
a,b
(f(X)) =
n
∑
i=0
c
i
(aX +b)
i
= c
n
a
n
X
n
+ terms of lower degrees in X.
Since c
n
a
n
̸= 0, this shows that deg ψ
a,b
(f(X)) = deg f(X).
(ii) Suppose that ψ
a,b
(p(X))  g(X)h(X) for g(X), h(X) ∈ k[X]. Choose k(X) ∈ k[X] so that
g(X)h(X) = k(X)ψ
a,b
(p(X)). Since ψ
a,b
is an isomorphism, we have
ψ
−1
a,b
(g(X))ψ
−1
a,b
(h(X)) = ψ
−1
a,b
(k(X))p(X)
104
and as p(X) is prime, p(X)  ψ
−1
a,b
(g(X)) or p(X)  ψ
−1
a,b
(h(X)). Hence ψ
a,b
(p(X))  g(X) or
ψ
a,b
(p(X))  h(X) and so ψ
a,b
(p(X)) is prime.
(iii) This follows from (ii) and Proposition 1.30.
1.9. (a) Addition and multiplication are given by the usual formulae
(
∞
∑
k=0
a
k
X
k
) + (
∞
∑
k=0
b
k
X
k
) =
∞
∑
k=0
(a
k
+b
k
)X
k
, (
∞
∑
k=0
a
k
X
k
)(
∞
∑
k=0
b
k
X
k
) =
∞
∑
k=0
(
k
∑
ℓ=0
a
ℓ
b
k−ℓ
)X
k
.
Clearly k[X] ⊆ k[[X]] is a subring. Given two nonzero elements a, b ∈ k[[X]] we may write
a =
∞
∑
k=k
0
a
k
X
k
, b =
∞
∑
ℓ
0
b
ℓ
X
ℓ
with a
k
0
̸= 0 ̸= b
ℓ
0
. Then the lowest degree term in ab is a
k
0
b
ℓ
0
X
k
0
+ℓ
0
with a
k
0
b
ℓ
0
̸= 0. Hence
ab ̸= 0. So k[[X]] is an integral domain.
(b) Let a =
∑
∞
k=0
a
k
X
k
∈ k[[X]]. Then a has an inverse in k[[X]] only if there is a b =
∑
∞
k=0
b
ℓ
X
ℓ
∈ k[[X]] with ab = 1, in particular this forces a
0
̸= 0 since otherwise the lowest term
in X in ab would be of degree greater than 0. Conversely, if a
0
̸= 0, then we can inductively
solve the system of equations
a
0
b
0
= 1,
n
∑
ℓ=0
a
ℓ
b
n−ℓ
= a
0
b
n
+a
1
b
n−1
+· · · +a
n
b
n
= 0 (n 1),
to ensure that ab = 1.
(c) We can deﬁne make the set k((X)) of all such ﬁnite tailed Laurent series into a ring with
addition and multiplication deﬁned by
(
∞
∑
k=k
1
a
k
X
k
) + (
∞
∑
k=k
2
b
k
X
k
) =
∞
∑
k=min{k
1
,k
2
}
(a
k
+b
k
)X
k
,
(
∞
∑
k=k
0
a
k
X
k
)(
∞
∑
ℓ=ℓ
0
b
ℓ
X
ℓ
) =
∞
∑
k=min{k
0
,ℓ
0
}
(
k
∑
j=0
a
ℓ
b
k−j
)X
k
.
Clearly k[[X]] ⊆ k((X)) is a subring. Notice that every element
∑
∞
k=k
0
a
k
X
k
∈ k((X)) with
k
0
< 0 can be written as
(
∑
r=0
a
r+k
0
X
r
)X
k
0
.
The inclusion inc: k[[X]] −→k((X)) extends to the monomorphism inc
∗
: Fr(k[[X]]) −→k((X))
for which
inc
∗
∑
∞
r=0
a
r+k
0
X
r
X
−k
0
= (
∞
∑
r=0
a
r+k
0
X
r
)X
k
0
,
so inc
∗
is surjective.
1.10. Here f(X) = (3X −3)d(X) + (−9X + 7).
1.11. Here f(X) = −X
3
−X
2
+X + 1 and d(X) = −X
3
−X with
f(X) = d(X) + (−X −X
2
+ 1) = d(X) + (2X
2
+ 2X + 1).
1.12. The reduction modulo p function
ρ: Z[X] −→F
p
[X]; ρ(f(X)) = f(X),
105
is a ring homomorphism. If f(X) = g(X)h(X) with g(X), h(X) ∈ Z[X], deg g(X) < deg f(X)
and deg h(X) < deg f(X), then
f(X) = ρ(g(X)h(X)) = ρ(g(X))ρ(h(X)) = g(X)h(X),
where deg g(X) < deg f(X) = deg f(X) and deg h(X) < deg f(X) = deg f(X). But this is
impossible since f(X) is irreducible. So f(X) must be irreducible.
X
3
−X +1 reduces modulo 3 to an irreducible since it has no roots modulo 3. So X
3
−X +1
is irreducible.
X
3
+2X +1 ≡ X
3
−X +1 (mod 3) so this polynomial reduces modulo 3 to an irreducible and
so is irreducible.
X
3
+X −1 reduces modulo 2 to an irreducible since it has no roots modulo 2. So X
3
+X −1
is irreducible.
X
5
−X + 1 is irreducible modulo 3 and 5 so is itself irreducible.
X
5
+ X − 1 = (X
3
+ X
2
− 1)(X
2
− X + 1) and 5X
3
− 10X + X
2
− 2 = (5X + 1)(X
2
− 2) so
neither of these is irreducible.
1.13. I
1
= (X
2
+1), I
2
= (X
2
+2), I
3
= (X
2
−2), I
4
= (X−
√
2), I
5
= (X
2
+2), I
6
= X
2
+X+1.
1.14. The image is
ε
√
2
Q[X] = Q[
√
2] = {a +b
√
2 : a, b ∈ Q}.
The image of ε
−
√
2
is ε
−
√
2
Q[X] = Q[
√
2] = ε
√
2
Q[X]. We have
ker ε
√
2
= ker ε
−
√
2
= (X
2
−2) ▹ Q[X]
which is a maximal ideal.
1.15. Notice that ω = (−1 +
√
3i)/2 = ζ
3
is a primitive 3rd root of unity and is a root of the
irreducible polynomial X
2
+X + 1 ∈ Q[X]. Then
ε
ω
Q[X] = Q[ω] = {a +bω : a, b ∈ Q}, ker ε
ω
= (X
2
+X + 1) ▹ Q[X],
where (X
2
+ X + 1) ▹ Q[X] is a maximal ideal. The other complex root of X
2
+ X + 1 is ω
2
,
so the evaluation homomorphism ε
ω
2 has ε
ω
2 Q[X] = ε
ω
Q[X] and ker ε
ω
2 = ker ε
ω
.
1.16. We have
ε
α
Q[X] = Q[α] = {a +bα +cα
2
+dα
3
: a, b, c, d ∈ Q}, ker ε
α
= (X
4
−2) ▹ Q[X],
and the latter ideal is maximal. The other complex roots of X
4
− 2 are −α, αi, −αi (notice
that two of these are real while the other two are not). Then
ker ε
−α
= ker ε
αi
= ker ε
−αi
= (X
4
−2) ▹ Q[X]
but although ε
−α
Q[X] = Q[α], we have
ε
αi
Q[X] = ε
−αi
Q[X] = Q[αi] = {a +bαi +cα
2
+dα
3
i : a, b, c, d ∈ Q} ̸= Q[α],
so ε
αi
Q[X] ̸= ε
α
Q[X] since one of these is a subset of R but the other is not.
If we replace Q by R, then in R[X],
X
4
−2 = (X
2
−
√
2)(X
2
+
√
2) = (X −
4
√
2)(X +
4
√
2)(X
2
+
√
2).
Let α be a root of X
4
−2. If α =
4
√
2, then
ε
α
R[X] = R[α] = {a +bα : a, b ∈ R} = R, ker ε
α
= (X −
4
√
2) ▹ R[X].
106
Similarly, if α = −
4
√
2, then
ε
−α
R[X] = R[−α] = {a −bα : a, b ∈ R} = R, ker ε
−α
= (X +
4
√
2) ▹ R[X].
If α
2
+ 2 = 0, then α / ∈ R and
ε
α
R[X] = R[α] = {a +bα : a, b ∈ R} = C, ker ε
α
= (X
2
+ 2) ▹ R[X].
1.17. First change variable to obtain
g(X) = f(X + 3) = X
3
−6X + 4.
Using Cardan’s method we have to solve the quadratic equation
U
2
+ 4U + 8 = 0,
which has roots
−2 ±2i = (
√
2)
3
e
3πi/4
.
Thus we can take
u =
√
2e
πi/4
ω
r
=
√
2
√
2
(1 +i)ω
r
= (1 +i)ω
r
(r = 0, 1, 3).
For the roots of g(X) we obtain 2,
√
3−1, −
√
3−1, while for f(X) we have 5,
√
3+2, −
√
3+2.
1.18. Work backwards with Cardan’s method. For α, take
−
q
2
= 10,
27q
2
+ 4p
3
108
= 108,
so q = −20 and p = 6. Thus α is a real root of f(X) = X
3
+ 6X − 20. Notice that 2 is a real
root of this polynomial and
f(X) = (X −2)(X
2
+ 2X + 10),
where X
2
+ 2X + 10 has no real roots. Therefore α = 2.
For β, take
−
q
2
= 1,
27q
2
+ 4p
3
108
=
28
27
,
so q = −2 and p = 1. Thus β is a real root of g(X) = X
3
+X−2 for which 1 is also a root and
g(X) = (X −1)(X
2
+X + 2),
where X
2
+X + 2 has no real roots. Therefore β = 1.
1.19. To see that the homomorphism
Aﬀ
1
(k) −→ Aut
k
(k[X]); A −→ α
A
−1,
described in the Proof of Example 1.60 is surjective, suppose that φ ∈ Aut
k
(k[X]) is any
automorphism. Let
φ(X) = a
0
+a
1
X +· · · +a
n
X
n
with a
i
∈ k and a
n
̸= 0. If n = 0 then φk[X] = k ⊆ k[X] so φ would not be surjective, hence
we must have n 1. Suppose that show that n > 1. Then
φk[X] = {c + 0 +c
1
φ(X) +· · · +c
k
φ(X)
k
: c
0
, c
1
, . . . , c
k
∈ k} = k[X].
107
But if k > 0 and c
k
̸= 0 then deg(c + 0 + c
1
φ(X) + · · · + c
k
φ(X)
k
) = kn > 1, so X / ∈ φk[X],
which gives a contradiction. So we must have n = 1. Therefore φ(X) = a
0
+a
1
X and so φ = α
A
for some A ∈ Aﬀ
1
(k).
1.20. Calculation.
1.21. We have
deg Φ
20
(X) = φ(20) = φ(4)φ(5) = 2 ×4 = 8
and
X
20
−1 = (X
10
−1)(X
10
+ 1) = (X
10
−1)(X
2
+ 1)(X
8
−X
6
+X
4
−X
2
+ 1).
Since the roots of X
10
−1 are the 10th roots of unity, we ﬁnd that
Φ
20
(X)  (X
2
+ 1)(X
8
−X
6
+X
4
−X
2
+ 1);
since cyclotomic polynomials are irreducible, we must have Φ
20
(X) = X
8
−X
6
+X
4
−X
2
+1.
1.22. (a) We have
X
p
k
−1 = (X
p
k−1
)
p
−1 = (X
p
k−1
−1)Φ
p
(X
p
k−1
),
so by (1.5),
∏
0jk
Φ
p
j (X) = Φ
p
(X
p
k−1
)
∏
0jk−1
Φ
p
j (X),
and therefore Φ
p
k(X) = Φ
p
(X
p
k−1
). The complex roots of Φ
p
(X) are the primitive pth roots
of 1, so the roots of Φ
p
k(X) are their p
k−1
st roots which are the primitive p
k
th roots of 1.
(b) Using the formula of Equation 1.4, we have
Φ
p
k(X) = Φ
p
(X
p
k−1
) = (X
p
k−1
−1)
p−1
+c
p−2
(X
p
k−1
−1)
p−2
+· · · +c
1
(X
p
k−1
−1) +c
0
,
where c
r
≡ 0 (mod p) and c
0
= p. The Idiot’s Binomial Theorem gives
X
p
k−1
−1 ≡ (X −1)
p
k−1
(mod p)
so
Φ
p
k(X) = (X −1)
(p−1)p
k−1
+c
′
p−2
(X −1)
(p−2)p
k−1
+· · · +c
′
1
(X −1)
p
k−1
+c
′
0
,
where c
′
r
≡ 0 (mod p). In fact,
c
′
0
= Φ
p
k(1) = Φ
p
(1) = c
0
= p,
so the Eisenstein Test can be applied to show that Φ
p
k(X) is irreducible over Q.
(c) First notice that
deg Φ
n
(X) = φ(n) = (p
1
−1) · · · (p
k
−1)p
r
1
−1
1
· · · p
r
k
−1
k
,
and
deg Φ
p
1
···p
k
(X
p
r
1
−1
1
···p
r
k
−1
k
) = φ(p
1
· · · p
k
)p
r
1
−1
1
· · · p
r
k
−1
k
= (p
1
−1) · · · (p
k
−1)p
r
1
−1
1
· · · p
r
k
−1
k
,
so deg Φ
n
(X) = deg Φ
p
1
···p
k
(X
p
r
1
−1
1
···p
r
k
−1
k
). Also, each root ξ of Φ
n
(X),
(ξ
p
r
1
−1
1
···p
r
k
−1
k
)
p
1
···p
k
= ξ
n
= 1,
and no smaller power of (ξ
p
r
1
−1
1
···p
r
k
−1
k
) has this property, hence (ξ
p
r
1
−1
1
···p
r
k
−1
k
) is a root of
Φ
p
1
···p
k
(X). This shows that Φ
n
(X)  Φ
p
1
···p
k
(X
p
r
1
−1
1
···p
r
k
−1
k
). As these are monic polynomials of
the same degree they are equal.
108
1.23. By Theorem 1.43, Φ
n
(X) =
∏
t=1,...,n−1
gcd(t,n)=1
(X −ζ
t
n
), so
Φ
n
(X
−1
) =
∏
t=1,...,n−1
gcd(t,n)=1
(X
−1
−ζ
t
n
)
= X
−φ(n)
∏
t=1,...,n−1
gcd(t,n)=1
(1 −Xζ
t
n
)
= X
−φ(n)
∏
t=1,...,n−1
gcd(t,n)=1
(1 −Xζ
n−t
n
)
= X
−φ(n)
∏
t=1,...,n−1
gcd(t,n)=1
(1 −Xζ
−t
n
)
= X
−φ(n)
∏
t=1,...,n−1
gcd(t,n)=1
(ζ
t
n
−X)
= (−1)
φ(n)
X
−φ(n)
∏
t=1,...,n−1
gcd(t,n)=1
(X −ζ
t
n
)
= (−1)
φ(n)
X
−φ(n)
Φ
n
(X).
Since 2  φ(n) when n > 2 and the result is immediate when n = 2, we see that desired equation
always holds.
1.24. We have
ζ
n
+ζ
−1
n
= e
2πi/n
+e
−2πi/n
= (cos(2πi/n) + sin(2πi/n) i) + (cos(2πi/n) −sin(2πi/n) i) = 2 cos(2πi/n).
Now we have
ζ
5
+ζ
−1
5
= 2 cos(2π/5), ζ
2
5
+ζ
−2
5
= (ζ
5
+ζ
−1
5
)
2
−2 = 4 cos
2
(2π/5) −2.
We also have Φ
5
(X) = X
4
+X
3
+X
2
+X + 1, so
ζ
4
5
+ζ
3
5
+ζ
2
5
+ζ
5
+ 1 = 0.
Rearranging and using the formulae ζ
4
5
= ζ
−1
5
, ζ
3
5
= ζ
−2
5
, we have
(ζ
2
5
+ζ
−2
5
) + (ζ
5
+ζ
−1
5
) + 1 = 0,
hence
4 cos
2
(2π/5) + 2 cos(2π/5) −1 = 0.
Thus a suitable polynomial is 4X
2
+ 2X −1 ∈ Q[X].
1.25. (a) In K[X], by the Idiot’s Binomial Theorem 1.10,
X
p
−1 = X
p
+ (−1)
p
= (X + (−1))
p
= (X −1)
p
.
By the Unique Factorization Property 1.33, the only root of this polynomial in K must be 1.
Similarly,
X
np
m
−1 = (X
n
−1)
p
m
109
and the only roots of this must be nth roots of 1.
(b) If u ∈ K is a root of this polynomial then u
p
= a. As in (a) we have
X
p
−a = X
p
−u
p
= (X −u)
p
,
so u is the only root in K.
Chapter 2
2.1. This is similar to Example 2.4.
2.2. It is obvious that [Q(
√
p,
√
q) : Q(
√
p)] 2; if [Q(
√
p,
√
q) : Q(
√
p)] = 1 then
√
q ∈ Q(
√
p),
say
√
q = a +b
√
p for some a, b ∈ Q. Then
q = (a +b
√
p)
2
= (a
2
+b
2
p) + 2ab
√
p,
giving the simultaneous pair of equations
a
2
+b
2
p = q, 2ab = 0.
If b = 0 then
√
q ∈ Q which contradicts the result of Qu. 2.1. If a = 0 then
√
q = b
√
p. Writing
b = b
1
/b
2
with b
1
, b
2
∈ Z and gcd(b
1
, b
2
) = 1, we obtain
b
2
2
q = b
2
1
p
and so p  b
2
and q  b
1
. Writing b
1
= b
′
1
q and b
2
= b
′
2
q for suitable b
′
1
, b
′
2
∈ Z, we obtain
(b
′
2
)
2
p
2
q = (b
′
1
)
2
q
2
p,
hence
(b
′
2
)
2
p = (b
′
1
)
2
q.
From this we obtain p  b
′
1
and q  b
′
2
; but then p  b
1
as well as p  b
2
, contradicting the fact
that gcd(b
1
, b
2
) = 1. So
√
q / ∈ Q(
√
p).
2.3. Arrange the induction carefully.
2.4. Notice that if v = ±u then b = v
2
= u
2
= a which is impossible; so v ̸= ±u. Then
u −v =
(u −v)(u +v)
u +v
=
u
2
−v
2
u +v
=
a −b
u +v
∈ K(u +v).
Hence
u =
1
2
((u +v) + (u −v)) ∈ K(u +v), v =
1
2
((u +v) −(u −v)) ∈ K(u +v).
So K(u, v) K(u +v) K(u, v) and therefore K(u +v) = K(u, v).
2.5. Since 1, i span the Qvector space Q(i), we have [Q(i) : Q] 2. But also if x, y ∈ R, then
x +yi = 0 ⇐⇒ x = y = 0, so 1, i is a basis for Q(i) over Q. Hence [Q(i) : Q] = 2.
2.6. First notice that [Q(
√
3) : Q] = 2 (with Qbasis 1,
√
3) and Q(
√
3) R. Also, i / ∈ Q(
√
3)
and since i
2
+ 1 = 0, Q(
√
3, i) = Q(
√
3)(i) has [Q(
√
3, i) : Q(
√
3)] = 2. By Theorem 2.6(ii),
[Q(
√
3, i) : Q] = [Q(
√
3, i) : Q(
√
3)] [Q(
√
3) : Q] = 2 ×2 = 4.
110
The following three subﬁelds of Q(
√
3, i) are distinct and are extensions of Q having degree 2:
L
1
= Q(
√
3), L
2
= Q(i), L
3
= Q(
√
3 i). Then [L
r
∩ L
s
: Q] > 1 ⇐⇒ L
r
∩ L
s
= L
r
= L
s
, so
L
r
∩ L
s
= Q whenever r ̸= s. The only real subﬁeld amongst these is L
1
.
C
2
∞ R
∞ Q(
√
3, i)
2
2
2
Q(
√
3)
2
Q(i)
2
Q(
√
3 i)
2
.
.
.
.
.
.
.
.
.
.
.
Q
2.7. (a) Since 5 is a prime,
[Q(ζ
5
) : Q] = [Q[X]/(Φ
5
(X)) : Q] = φ(5) = 5 −1 = 4.
(b) We have ζ
5
= cos(2π/5) +sin(2π/5) i ∈ Q(ζ
5
). But also ζ
−1
5
∈ Q(ζ
5
) and ζ
−1
5
= cos(2π/5) −
sin(2π/5) i ∈ Q(ζ
5
). Hence we have
cos(2π/5) =
1
2
ζ
5
+ζ
−1
5
∈ Q(ζ
5
), sin(2π/5) i =
1
2
ζ
5
−ζ
−1
5
∈ Q(ζ
5
).
(c) This can be found by repeated use of the double angle formula
cos(A+B) = cos Acos B −sin Asin B.
The polynomial T
n
(X) expressing cos nθ in terms of cos θ is called the nth Chebyshev polyno
mial, see Remark 6.6.
(d) For k = 0, 1, 2, 3, 4, cos(5(2kπ/5)) = cos(2kπ) = 1, so T
5
(cos 2kπ/5) − 1 = 0. So each of
the numbers cos(2kπ/5) is a root of the polynomial T
5
(X) −1 = (X −1)(4X
2
+2X −1)
2
. For
k = 1, 2, 3, 4, cos(2kπ/5) is a root of 4X
2
+ 2X −1, therefore
Q(cos(2π/5))
∼
= Q[X]/(4X
2
+ 2X −1), [Q(cos(2kπ/5)) : Q] = 2.
(e) Q(ζ
5
)
2
Q(cos(2π/5))
2
Q
2.8. This is similar to the previous question.
2.9. (a) If α ∈ Aut
Q
(E
n
) then α(2
1/n
)
n
= α(2) = 2, so α(2
1/n
) ∈ E
n
is also a real n
th root of 1. If n is odd, the only possibility is α(2
1/n
) = 2
1/n
, so α = id. If n is even, the
possibilities are α(2
1/n
) = ±2
1/n
. We can realize this automorphism starting with the evaluation
111
homomorphism ε
2
1/n : Q[X] −→ E
n
and precomposing with the isomorphism ψ: Q[X] −→Q[X]
for which ψ(X) = −X to form ε
′
2
1/n
= ε
2
1/n ◦ ψ. On passing to the quotient homomorphism of
ε
′
2
1/n
we obtain an automorphism τ
n
of E
n
under which τ
n
(2
1/n
) = −2
1/n
.
(b) Since E R, an automorphism α ∈ Aut
Q
(E) has the eﬀect
α(2
1/n
) =
2
1/n
if n is odd,
±2
1/n
if n is even.
If for some n we have α(2
1/n
) = −2
1/n
then
−2
1/n
= α(2
1/n
) = α(2
1/2n
)
2
> 0
since α(2
1/2n
) ∈ R. This contradiction shows that α(2
1/n
) = 2
1/n
for every n, so α = id.
(c) Assuming there are only 6 such subﬁelds, they form the following tower.
E
12
3
2
E
4
2
E
6
3
2
E
2
2
E
3
3
Q
(d) This element is a root of the polynomial
(X −(2
1/2
+ 2
1/3
))(X −(−2
1/2
+ 2
1/3
)) = X
2
−2(2
1/3
)X + 2
2/3
−2 ∈ E
3
[X],
so it is certainly an element of E
6
which is the only degree 2 extension of E
3
. If 2
1/2
+2
1/3
∈ E
3
then 2
1/2
∈ E
3
, which would imply 2 = [E
2
: Q]  [E
3
: Q] = 3 which is false, so 2
1/2
+2
1/3
/ ∈ E
3
;
a similar argument shows that 2
1/2
+ 2
1/3
/ ∈ E
2
. Writing ω = e
2πi/3
, 2
1/2
+ 2
1/3
is a root of
(X −(2
1/2
+ 2
1/3
))(X −(2
1/2
+ 2
1/3
ω))(X −(2
1/2
+ 2
1/3
ω
2
))
= X
3
−3(2
1/2
)X
2
+ 6X −(2 + 2(2
1/2
)) ∈ E
2
[X],
so it cannot lie in E
4
since 2
1/2
+ 2
1/3
/ ∈ E
2
and 3 [E
4
: E
2
] = 2. So 2
1/2
+ 2
1/3
is in E
6
and
E
12
and none of the others.
Chapter 3
3.1. Clearly, t is algebraic over K if and only if ker ε
t
̸= (0), i.e., (i) ⇐⇒ (ii). By Theorem 2.9,
(ii) ⇐⇒ (iii). Hence these three conditions are indeed equivalent.
3.2. The diagrams at the bottom indicate useful subﬁelds of the splitting ﬁelds occurring in
each of these examples.
p
1
(X) = X
4
−X
2
+1: The polynomial X
2
−X +1 has the complex roots e
±πi/3
=
1 ±
√
3 i
2
, so
the four roots of p
1
(X) are the complex square roots of these numbers, i.e., ±e
±πi/6
. Explicitly
these are
√
3
2
+
1
2
i, −
√
3
2
−
1
2
i,
√
3
2
−
1
2
i, −
√
3
2
+
1
2
i.
112
The splitting ﬁeld is E = Q(
√
3, i) and [E : Q] = 4.
p
2
(X) = X
6
− 2: The roots are the six complex 6th roots of 2, i.e.,
6
√
2e
2kπi/6
=
6
√
2e
kπi/3
for
k = 0, 1, 2, 3, 4, 5. Explicitly, these are
6
√
2,
6
√
2
2
+
6
√
2
√
3
2
i, −
6
√
2
2
+
6
√
2
√
3
2
i, −
6
√
2, −
6
√
2
2
−
6
√
2
√
3
2
i,
6
√
2
2
−
6
√
2
√
3
2
i.
The splitting ﬁeld is E = Q(
6
√
2,
√
3i) = Q(
6
√
2)(
√
3i) which has degree [E : Q] = 12.
p
3
(X) = X
4
+ 2: The roots are the four 4th roots of −2, i.e.,
4
√
2e
(2k+1)πi/4
for k = 0, 1, 2, 3.
Explicitly these are
1
4
√
2
+
1
4
√
2
i, −
1
4
√
2
+
1
4
√
2
i, −
1
4
√
2
−
1
4
√
2
i,
1
4
√
2
−
1
4
√
2
i.
The splitting ﬁeld is E = Q(
4
√
2, i) and [E : Q] = 8.
p
4
(X) = X
4
+ 5X
3
+ 10X
2
+ 10X + 5: Notice that
p
4
(Y −1) = Y
4
+Y
3
+Y
2
+Y + 1 = Φ
5
(Y ),
so the splitting ﬁeld of p
4
(X) over Q is the same as that of Φ
5
(Y ) over Q and this is the
cyclotomic ﬁeld Q(ζ
5
) where ζ
5
= cos(2π/5) + sin(2π/5)i with [Q(ζ
5
) : Q] = 4; in fact we have
Q(ζ
5
) = Q(cos(2π/5), sin(2π/5)i).
Q(
√
3, i)
Q(
√
3)
2
Q
2
Q(
6
√
2,
√
3, i)
Q(
6
√
2,
√
3)
2
Q(
√
3)
6
Q
2
Q(
4
√
2, i)
Q(
4
√
2)
2
Q
4
Q(cos(2π/5), sin(2π/5)i)
Q(cos(2π/5))
2
Q
2
3.3. List the three roots of X
3
−2 as u
1
=
3
√
2, u
2
=
3
√
2ζ
3
, u
3
=
3
√
2ζ
2
3
. Then each automorphism
α ∈ Aut
Q
(Q(
3
√
2, ζ
3
)) permutes these roots, so can be identiﬁed with the unique permutation
σ
α
∈ S
3
for which
α(u
i
) = u
σ
α
(i)
(i = 1, 2, 3).
We ﬁnd that (using cycle notation for permutations)
σ
id
= id, σ
α
0
= (2 3), σ
α
1
= (1 2 3), σ
α
′
1
= (1 2), σ
α
2
= (1 3 2), σ
α
′
2
= (1 3).
These are the six elements of S
3
, therefore Aut
Q
(Q(
3
√
2, ζ
3
))
∼
= S
3
.
3.4. Irreducibility is a consequence of the polynomial version of the Eisenstein Test 1.48.
Suppose that t ∈ k(T) is a root of g(X); then using the Idiot’s Binomial Theorem we have
(X −t)
p
= X
p
−t
p
= X
p
−T,
so t is in fact a root of multiplicity p, hence it is the only root of g(X) in k(T). This also gives
the factorization of g(X) into linear factors over k(T).
3.5. Q(
√
5,
√
10)/Q: Here [Q(
√
5,
√
10) : Q] = 4 and the element
√
5 +
√
10 has degree 4 with
minimal polynomial X
4
−30X
2
+ 25 which has roots ±
√
5 ±
√
10.
113
Q(
√
2, i)/Q: Here [Q(
√
2, i) : Q] = 4 and the element
√
2 + i has degree 4 with minimal
polynomial X
4
−2X
2
+ 9 which has roots ±
√
2 ±i.
Q(
√
3, i)/Q: Here [Q(
√
3, i) : Q] = 4 and the element
√
3 + i has degree 4 with minimal
polynomial X
4
−4X
2
+ 16 which has roots ±
√
3 ±i.
Q(
4
√
3, i)/Q: Here [Q(
4
√
3, i) : Q] = 8 and the element
4
√
3 + i has degree 8 with minimal
polynomial X
8
+ 4X
6
+ 40X
2
+ 4 which has roots ±
4
√
3 ±i and ±
4
√
3i ±i.
3.6. The induction is straightforward. Here is the argument that K(u, v)/K is simple. We
assume that K is inﬁnite since otherwise the result will be proved in Proposition 5.16.
Consider the subﬁelds K(u + tv) K(u, v) with t ∈ K. Then there are only ﬁnitely many
of these, so there must be s, t ∈ K such that s ̸= t and K(u +sv) = K(u +tv). Then
(s −t)v = (u +sv) −(u +tv) ∈ K(u +tv),
hence v ∈ K(u +tv). This implies that
u = (u +tv) −tv ∈ K(u +tv),
hence K(u, v) K(u +tv) K(u, v) and so K(u, v) = K(u +tv).
3.7. If E/K is a quadratic extension then for any u ∈ E − K we have 1 < [K(u) : K] 2,
so [K(u) : K] = 2 = [E : K] and therefore K(u) = E. Then minpoly
K,u
(X) must factor into
linear factors over E, so both its roots in K lie in E. This shows that E is normal over K.
The example F
2
(Z)/F
2
(Z
2
) is not separable since X
2
− Z
2
∈ F
2
(Z
2
)[X] is irreducible but
not separable (see Qu. 3.4). If char K ̸= 2 then all quadratic polynomials over K are separable.
3.8. Let E C be a splitting subﬁeld for f(X) over Q. Then if v ∈ C is a nonreal root of
f(X) we have v / ∈ Q(u), so f(X) does not split over Q(u) even though it has a root in this ﬁeld.
This means that there is a monomorphism φ ∈ Mono
Q
(Q(u), C) = Mono
Q
(Q(u), Q) for which
φ(u) = v, hence φQ(u) ̸= Q(u) and so Q(u)/Q is not normal.
Chapter 4
4.1. By Theorem 3.81 we know that splitting ﬁelds are always normal, so it is only necessary
to show that the splitting ﬁeld E of p(X) over K is separable over K. Since E is obtained by
repeatedly adjoining roots of p(X), the result follows from Proposition 3.73 together with the
fact that if L/K E/K is separable and v ∈ E is a root of p(X), then L(v)/K is separable.
4.2. (a) Suppose that f(X) = c
3
X
3
+c
2
X
2
+c
1
X +c
4
with c
3
̸= 0. Then
f(uX +v) =
c
3
u
3
X
3
+ (3c
3
vu
2
+c
2
u
2
)X
2
+ (3c
3
uv
2
+c
1
u + 2c
2
uv)X + (c
3
v
3
+c
4
+c
1
v +c
2
v
2
),
so if we take u to be any cube root of c
3
and u = −c
2
/3c
3
then f(uX +v) has the desired form.
Notice that v ∈ K(u) and then f(uX +v) ∈ K(u), so provided that we can ﬁnd a cube root of
1/c
3
in K, we have f(uX +v) ∈ K.
(b) Viewing Gal(E/K) as a subgroup of S
3
, by Theorem 4.8 we know that 3 divides  Gal(E/K);
but the only subgroups of S
3
with this property are S
3
and A
3
.
(c) This is a tedious calculation! See Section 4.7 for the rest of this question.
114
4.3. If a/b is a rational root of f(X), we may assume that gcd(a, b) = 1. Now a
3
−3ab
2
+b
3
=
0, which easily implies that a, b = ±1; but 1 is certainly not a root. Hence there are no
rational roots and so no proper rational factors. By the formula following Proposition 4.25, the
discriminant of f(X) is
∆ = −27 −4(−3)
3
= 81 = 9
2
.
If the distinct roots of f(X) in C are u, v, w, the splitting subﬁeld K(v, w) = Q(u, v, w) C
satisﬁes 3  [Q(u, v, w) : Q] and [Q(u, v, w) : Q]  3! = 6. The Galois group Gal(Q(u, v, w)/Q) is a
subgroup of S
3
(viewed as the permutation group of {u, v, w}). Since the discriminant is a square
in Q, Proposition 4.26 implies that Gal(Q(u, v, w)/Q) A
3
∼
= Z/3. So  Gal(Q(u, v, w))/Q) = 3
and Gal(Q(u, v, w)) is cyclic of order 3 whose generator is a 3cycle which cyclically permutes
u, v, w.
4.4. (a) This should be a familiar result.
(b) The centre of D
8
is
⟨
α
2
⟩
which has order 2, and there are three normal subgroups of order
4, namely
⟨α⟩ = {ι, α, α
2
, α
3
},
⟨
α
2
, β
⟩
= {ι, α
2
, β, βα
2
},
⟨
α
2
, βα
⟩
= {ι, α
2
, βα, βα
3
}.
Notice that there are also four nonnormal subgroups of order 2,
⟨β⟩ = {ι, β}, ⟨βα⟩ = {ι, βα},
⟨
βα
2
⟩
= {ι, βα
2
},
⟨
βα
3
⟩
= {ι, βα
3
}.
4.5. This is an example of case (iii) of Kaplansky’s Theorem and we use the notation of the
proof. The discriminant here is δ
2
= −12, so we can take δ = 2
√
3i. The roots of X
2
+ 3 are
±
√
3i, so we may assume
u =
4
√
3 ζ
8
=
√
2
4
√
3
2
(1 +i), v =
4
√
3 ζ
−1
8
=
√
2
4
√
3
2
(1 −i),
where as usual ζ
8
= e
2πi/8
= (1 +i)/
√
2. Hence we have uv =
√
3 and uvδ = 6i. This gives the
diagram of subﬁelds of E
E = Q(
4
√
3 ζ
8
,
4
√
3 ζ
−1
8
) = Q(
4
√
3, ζ
8
)
Q(
√
3, i)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Q(
√
3i)
2
Q(
√
3)
2
Q(i)
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Q
Then α is the restriction of complex conjugation to E, while β(
√
3i) =
√
3i and β(
√
3) = −
√
3,
hence also β(i) = −i. Using the choices of the proof, we have
β(
4
√
3 ζ
8
) = −
4
√
3 ζ
8
, β(
4
√
3 ζ
−1
8
) = β(−
4
√
3 ζ
8
i) = −
4
√
3 ζ
8
i.
The eﬀects of σ and γ on the four roots
4
√
3 ζ
8
,
4
√
3 ζ
−1
8
, −
4
√
3 ζ
8
, −
4
√
3 ζ
−1
8
of f(X) are given in
permutation notation by σ = (1 4 3 2) and α = (1 2)(3 4), and these generate a dihedral
115
subgroup of S
4
. Using the previous question (but beware that the notation there is inconsistent
with that of the present situation!) we have the normal subgroups
⟨
σ
2
⟩
, ⟨σ⟩ ,
⟨
σ
2
, α
⟩
,
⟨
σ
2
, ασ
⟩
,
and these have ﬁxed ﬁelds
E
⟨σ
2
⟩
= Q(
√
3, i), E
⟨σ⟩
= Q(i), E
⟨σ
2
,α⟩
= Q(
√
3), E
⟨σ
2
,ασ⟩
= Q(
√
3 i),
each of which is a normal extension of Q.
4.6. Q(X
3
− 10)/Q: This is similar to Example 4.20, with splitting ﬁeld Q(
3
√
10, ζ
3
) and
Gal(Q(
3
√
10, ζ
3
)/Q)
∼
= S
3
.
Q(
√
2)(X
3
−10)/Q(
√
2): The splitting ﬁeld is Q(
√
2,
3
√
10, ζ
3
), [Q(
√
2,
3
√
10) : Q(
√
2)] = 3 and
Q(
√
2,
3
√
10) Q(
√
2,
3
√
10, ζ
3
).
Since ζ
3
is not real, [Q(
√
2,
3
√
10, ζ
3
) : Q(
√
2)] = 6. The Galois group is isomorphic to S
3
.
Q(
√
3 i)(X
3
− 10)/Q(
√
3 i): Here Q(
√
3 i) = Q(ζ
3
), with [Q(ζ
3
) : Q] = 2. The splitting ﬁeld is
Q(
3
√
10, ζ
3
) and [Q(
3
√
10, ζ
3
) : Q(ζ
3
)] = 3, hence Gal(Q(
3
√
10, ζ
3
)/Q(ζ
3
))
∼
= Z/3 with generator σ
for which σ(
3
√
10) =
3
√
10 ζ
3
.
Q(
√
23 i)(X
3
−X −1)/Q(
√
23 i): First note that X
3
−X −1 ∈ Z[X] must be irreducible since
its reduction modulo 2, X
3
+ X + 1 ∈ F
2
[X], has no root in F
2
and hence has no linear factor
(see Qu. 1.10). To proceed further we can use the ideas of Qu. 4.2 above (see also Section 4.7).
The discriminant of the polynomial X
3
− X − 1 is ∆ = −23 and so δ =
√
23 i. Then if
E = Q(
√
23 i)(X
3
− X − 1) is the splitting ﬁeld of X
3
− X − 1 over Q, Gal(E/Q)
∼
= S
3
and
Gal(E/Q(
√
23 i))
∼
= A
3
.
K(X
3
− X − 1)/K for K = Q, Q(
√
5), Q(
√
5 i), Q(i): Continuing the preceding discussion,
notice that [E ∩ R : Q] = 3, so
√
5 / ∈ E, hence
Q(
√
5)(X
3
−X −1) = Q(X
3
−X −1)(
√
5)
and
[Q(
√
5)(X
3
−X −1) : Q(
√
5)] = [Q(X
3
−X −1) : Q] = 6,
hence Gal(Q(
√
5)(X
3
−X −1)/Q(
√
5))
∼
= S
3
. Similarly,
√
5 i / ∈ E and i / ∈ E, hence
Gal(Q(
√
5 i)(X
3
−X −1)/Q(
√
5 i))
∼
= S
3
∼
= Gal(Q(i)(X
3
−X −1)/Q(i)).
4.7. (a) Since char K ̸= 0, f
′
(X) = pX
p−1
̸= 0, so if u ∈ L is any root of f(X) then
f
′
(u) = pu
p−1
̸= 0. By Proposition 3.55, there are no multiple roots, hence p distinct roots. If
u, v ∈ L are distinct roots, then (vu
−1
)
p
= 1, so v = uξ for ξ ∈ K a pth root of 1 with ξ ̸= 1.
(b) If there is a root u / ∈ K, the Galois group Gal(L/K) acts in the following way. By The
orem 4.8, there must be an element γ ∈ Gal(L/K) with γ(u) ̸= u. We can write γ(u) = uξ
γ
where ξ
γ
̸= 1 is a pth root of 1. Since γ(ξ
γ
) = ξ
γ
, for r 1 we have γ
r
(u) = uξ
r
γ
, which can
only equal u if p  r. So u must have at least p conjugates which are all roots of f(X). Since
deg f(X)
p
, every root of f(X) is conjugate to u, so f(X) must be irreducible over K.
(c) Suppose that f(X) = g(X)h(X) with g(X) ∈ K[X] monic irreducible and 0 < d =
deg g(X) < p. Let L/K with L a splitting ﬁeld for f(X) over K and let w ∈ L be a root
of g(X). Arguing as in (a), we know that each root of g(X) has the form wξ where ξ is some
116
pth root of 1; moreover, L must contain p distinct pth roots of 1. Now the constant coeﬃcient
of g(X) is g(0) = (−1)
d
ξ
0
w
d
∈ K where ξ
0
is a pth root of 1. So
g(0)
p
= (−1)
dp
ξ
p
0
(w
p
)
d
= (−1)
dp
a
d
,
from which it follows that a
d
is a pth power in K. As gcd(p, d) = 1, there are integers r, s such
that rp +sd = 1, so we have
a = (a
r
)
p
(a
d
)
s
= a pth power in K.
Hence if f(X) is not irreducible in K[X] it has a root in K.
4.8. If u ∈ L is a root of f(X) in an extension L/K then by the Idiot’s Binomial Theorem 1.10
X
p
−a = X
p
+ (−u)
p
= (X −u)
p
,
so u is the only such root in L and f(X) splits over L. If (X − u)
d
∈ K[X] for some d with
1 < d < p then u
d
∈ K. Since gcd(d, p) = 1, there are integers r, s such that rd+sp = 1. Hence
(u
d
)
s
(u
p
)
r
= u, where the left hand side is in K. This shows that u ∈ K. Hence either f(X)
has a root in K or it must be irreducible over K.
Chapter 5
5.1. By Theorem 1.17, an integral domain D always admits a monomorphism into a ﬁeld
j : D −→ F (e.g., F can be taken to be the ﬁeld of fractions of D), so any subgroup U D
×
becomes isomorphic to a subgroup jU F
×
, and if U is ﬁnite so is jU. Therefore jU and U
are cyclic.
5.2. The only root of X
2
+ 1 in F
2
is the multiple root 1.
5.3. The ﬁeld F
p
d[X]/(f(X)) is an extension of F
p
d which has degree n, hence it is a ﬁnite
ﬁeld with p
dn
elements, hence Proposition 5.6 implies that it is isomorphic to F
p
dn. Since the
extension F
p
dn/F
p
d is normal, F
p
dn is a splitting ﬁeld for f(X) over F
p
d.
5.4. (a) Here 41 is prime. Since 8  (41 −1), there is a primitive 8th root of unity in F
41
. 6 is
a primitive root for F
41
and 6
5
≡ 27 (mod 4)1 has order 8.
(b) Here 5 is prime 4  (5 −1), so there is a primitive 4th root of unity in F
×
5
, but no primitive
8th root of unity. In fact, 2 and 3 have order 4, so these are primitive roots for F
5
. Notice that
in F
5
[X],
X
8
−1 = (X
4
−1)(X
4
+ 1) = (X
4
−1)(X
2
−2)(X
2
−3),
where the polynomials X
2
−2 and X
2
−3 are irreducible. Therefore F
25
is the splitting ﬁeld for
X
8
−1 over F
5
and we have F
25
∼
= F
5
(u) = F
5
(v), where u
2
= 2 and v
2
= 3, so ±u and ±v are
primitive 8th roots of unity. To ﬁnd an element of order 24 in F
×
25
, we ﬁrst ﬁnd one of order 3.
Consider the polynomial X
2
+X + 1 ∈ F
5
[X]; in F
5
, this has roots which have order 3. These
roots are given by (−1 ±w)/2, where w
2
= (1 −4) = −3 = 2, hence they are
(−1 ±u)
2
= −3 ±3u.
Now the elements ±(2 ±2u)u = ±(±4 + 2u) = ±4 ±2u all have order 8 ×3 = 24.
(c) Here 11 is prime and 8  (121−1) = 120, so F
121
is the splitting ﬁeld of X
8
−1 over F
11
. The
polynomial X
2
+1 is irreducible over F
11
so F
121
= F
11
(z) where z
2
= −1. Since 120 = 8×3×5,
it is suﬃcient to ﬁnd elements of order 8, 3 and 5 whose product will have order 120.
117
Suppose that a+bz ∈ F
121
with a, b ∈ F
11
. If this element has order 8, then (a+bz)
2
= ±z.
So let us solve
(a
2
−b
2
) + 2abz = z.
Then 2ab = 1 and b
2
= a
2
, hence b = ±a. Now we have 2a
2
= ±1 and so a
2
= ±1/2 = ±6.
Now 6 is not a square in F
11
but
7
2
≡ −6 ≡ 4
2
(mod 11),
so we have a = 4, b = ±4 and a = 7, b = ±7. Therefore the elements of order 8 in F
×
121
are
4 ±4z and 7 ±7z.
By the same approach as in (b), the elements of order 3 in F
121
are (−1 ±5z)/2 = 5 ±8z.
2 is a primitive root for F
11
so 4 = 2
2
has order 5.
Combining these we obtain the following primitive roots for F
121
: 7 ±z, 10 ±4z.
(d) In F
2
[X] we have X
8
−1 = (X −1)
8
, whose only root in F
2
is 1. So the splitting ﬁeld is F
2
.
5.5. Notice that F
p
(w) is a splitting ﬁeld of the separable polynomial X
p
d
−1
− 1 over F
p
, so
if w ∈ F
×
p
ℓ
then F
p
(w) F
p
ℓ . Since F
p
(w) = F
p
d we have d ℓ; we also have deg
F
p
w = d.
The number of conjugates of w is d, hence each primitive root of F
p
d has d conjugates and the
total number of these is the number of generators of the cyclic group F
×
p
d
∼
= Z/(p
d
− 1), i.e.,
φ(p
d
− 1). Hence d  φ(p
d
− 1). This can also be interpreted in terms of the evident action of
Gal(F
p
d/F
p
)
∼
= Z/d on the set of all primitive roots of F
p
d; each orbit has exactly d elements,
so the number of orbits is φ(p
d
−1)/d which is an integer.
5.6. (a) First note that g
′
a
(X) = −1, so g
a
(X) is separable, hence E/K is separable. If u ∈ E
is a root of g
a
(X), then for t ∈ F
p
d,
g
a
(u +t) = (u +t)
p
d
−(u +t) −a = (u
p
d
−u −a) + (t
p
d
−t) = (u
p
d
−u −a) = 0,
hence u+t is also a root of g
a
(X). This means that E = K(u) since all the other roots of g
a
(X)
lie in K(u). As g
a
(X) is irreducible over K, [E : K] = p
d
=  Gal(E/K) and so the following
p
d
automorphisms are the elements of Gal(E/K):
σ
t
: E −→ E; σ
t
(u) = u +t (t ∈ F
p
d).
It is easy to check that for s, t ∈ F
p
d, σ
s
◦σ
t
= σ
s+t
. Hence there is an isomorphism Gal(E/K)
∼
=
F
p
d with σ
t
corresponding to t ∈ F
p
d.
(b) If g
a
(X) is irreducible over K then it cannot have a root in K since its degree is greater
than 1.
Conversely, suppose that g
a
(X) has no root in K. Then if u ∈ E is any root of g
a
(X) in
a splitting ﬁeld over K, the other roots are the p elements u + t ∈ E (t ∈ F
p
). If u + t
0
̸= u
is a conjugate of u with 0 ̸= t
0
∈ F
p
, there must be an element τ
t
0
∈ Gal(E/K) for which
τ
t
0
(u) = u + t
0
. Then ⟨τ
t
0
⟩ must be isomorphic to a nontrivial subgroup of F
p
, but this must
be F
p
since this group is simple. Hence, u must have p conjugates and so g
a
(X) is irreducible
over K.
(c) If K is a ﬁnite ﬁeld and d > 1 then if g
a
(X) were irreducible over K, then by (a), E would
be ﬁnite and Gal(E/K)
∼
= F
p
d. But F
p
d is not cyclic, yet we know from Proposition 5.23 that
Gal(F
p
d/F
p
)
∼
= Z/d is cyclic.
118
5.7. (a) By Proposition 5.12, F
×
q
is a cyclic group. If p = 2 then F
×
2
d
 = 2
d
−1, which is odd,
so every element of F
×
2
d
is a square; we may therefore take λ
2
d(u) = 1 for all u ∈ F
×
2
d
. So now
suppose that p is odd. Then F
×
p
d
 = p
d
− 1, which is even. The set of squares in F
×
p
d
is the
normal subgroup
(F
×
p
d
)
2
= {u
2
: u ∈ F
×
p
d
} F
×
p
d
and it is easily seen that its quotient group has order 2, hence
F
×
p
d
/(F
×
p
d
)
2
∼
= {±1}.
We may use this group isomorphism to deﬁne λ
q
. Clearly we have
ker λ
q
= (F
×
p
d
)
2
.
λ
q
is surjective if and only if p is odd.
Remark: when d = 1, λ
p
(u) =
u
p
, the Legendre symbol of u from Number Theory.
(b) If u ∈ Σ
q
, then either u = 0 or u ̸= 0 and u = (±v)
2
for some v ∈ F
×
q
. Thus we have
Σ
q
 = 1 +
(q −1)
2
=
(q + 1)
2
.
Then
t −Σ
q
 = Σ
q
 =
(q + 1)
2
.
(c) Since Σ
q
∪ (t −Σ
q
) ⊆ F
q
, we have
q Σ
q
∪ (t −Σ
q
) = Σ
q
 +t −Σ
q
 −Σ
q
∩ (t −Σ
q
).
This implies that
q (q + 1) −Σ
q
∩ (t −Σ
q
)
and so
Σ
q
∩ (t −Σ
q
) 1.
Thus for every t ∈ F
q
, there are u, v ∈ F
q
(possibly 0) for which u
2
= t −v
2
, whence t = u
2
+v
2
.
(d) By (c), we may write −1 = a
2
+b
2
for some a, b ∈ F
q
, i.e.,
1
2
+a
2
+b
2
= 0.
Chapter 6
6.1. Now when n = 1, G
∼
= Z/p, which is abelian. Suppose that the result holds whenever
G = p
k
with k < n. Now if G = p
n
, recall that by Cauchy’s Lemma, the centre Z of G
is nontrivial. Hence G/Z has order G/Z = p
k
with k < n. By the inductive hypothesis,
there is a normal subgroup M ▹ G/Z with M = p
k−1
. By one of the Isomorphism Theorems,
there is a normal subgroup N ▹ G containing Z and satisfying N/Z = M ⊆ G/Z. Clearly
N = Z M = p
n−1
. This establishes the inductive step and hence the desired result.
6.2. In this situation, for any nonzero t ∈ K, −t ̸= t (since otherwise 2t = 0 and so t = 0).
If ζ ∈ K is a primitive nth root of unity, then (−ζ)
n
= (−1)
n
ζ
n
= −1, while (−ζ)
2n
=
(−1)
2n
ζ
2n
= 1. Hence −ζ ∈ K is a primitive 2nth root of unity.
119
6.3. Write n = 2
k
p
r
1
1
· · · p
r
s
s
, where each p
j
is an odd prime, p
1
< p
2
< · · · < p
s
, r
j
1 and
k 0. Then
φ(n) = φ(2
k
)φ(p
r
1
1
) · · · φ(p
r
s
s
) = φ(2
k
)(p
1
−1)p
r
1
−1
1
· · · (p
s
−1)p
r
s
−1
s
.
If s > 0 then φ(n)  4 happens precisely when r
1
= · · · = r
s
= 1 and one of the following
possibilities occurs:
• p
1
= 5, s = 1 and k = 0 (hence n = 5);
• p
1
= 3, s = 1 and k = 0, 1, 2 (hence n = 3, 6, 12);
• s = 0 and k = 0, 1, 3 (hence n = 1, 2, 4, 8).
Q(i): Here degree [Q(i) : Q] = 2 and clearly the four 4th roots of unity ±1, ±i lie in this ﬁeld.
As φ(5) = 4, it has no 5th roots of unity except 1. If it contained a 3rd root of unity then it
would contain
√
3 and so Q(
√
3, i) Q(i) which is impossible since [Q(
√
3, i) : Q] = 4. From
this we see that the only roots of unity in Q(i) are ±1, ±i.
Q(
√
2 i): This ﬁeld contains only the square roots of unity ±1.
Q(
√
3 i): This contains the six 6th roots of unity ±1, ±
1
2
±
√
3
2
i.
Q(
√
5 i): This ﬁeld contains only the square roots of unity ±1.
6.4. (a) We have φ(24) = φ(8)φ(3) = 4 × 2 = 8. The elements of Z/24 which are invertible
are the residue classes modulo 24 of the numbers 1, 5, 7, 11, 13, 17, 19, 23. For each of these
numbers r, the residue class modulo 24, r, satisﬁes r
2
= 1, hence these all have order 2 except
1 which has order 1. Since (Z/24)
×
is abelian, it is isomorphic to Z/2 ×Z/2 ×Z/2. The eﬀect
of these elements on Q(ζ
24
) is given by r · ζ
r
24
. Notice that 23 acts like complex conjugation.
The eﬀect on Q(cos(π/12)) is given by
r · cos(π/12) = cos(πr/12),
so in particular,
−r · cos(π/12) = cos(−πr/12) = cos(πr/12) = r · cos(π/12).
(b) This is similar to (a). We have φ(20) = φ(4)φ(5) = 2 ×4 = 8 and the elements of (Z/20)
×
are the residue classes modulo 20 of the numbers 1, 3, 7, 9, 11, 13, 17, 19. This time there are
elements of order 4, for instance 7 and 13. Then we have (Z/20)
× ∼
= Z/2 ×Z/4.
6.5. For any n 1, let ζ
n
= e
2πi/n
= cos(2π/n) + sin(2π/n) i. Notice that if n is odd, then
Q(ζ
n
) = Q(−ζ
n
) where −ζ
n
is a primitive 2nth root of unity, so we might as well assume that
n is even from now on. We also have
ζ
n
−ζ
−1
n
= 2 sin(2π/n) i ∈ Q(ζ
n
).
(a) If 4 n then writing n = 2k with k odd, we have
[Q(ζ
n
) : Q] = φ(2k) = φ(2)φ(k) = φ(k),
while
[Q(ζ
2n
) : Q] = φ(4k) = φ(4)φ(k) = 2φ(k).
120
Hence, Q(ζ
n
) cannot contain ζ
2n
and by another simple argument it cannot contain i = ζ
k
2n
. So
we see that sin(2π/n) / ∈ Q(ζ
n
) in this situation. Notice that since i = ζ
k
2n
,
sin(2π/n) =
ζ
2
2n
−ζ
−2
2n
2i
∈ Q(ζ
2n
),
and by Theorem 6.3,
sin(2π/n) ∈ Q(ζ
2n
) ∩ R = Q(cos(π/n)).
Also, we have
[Q(cos(π/n)) : Q] = 2[Q(cos(2π/n)) : Q],
hence
Q(cos(π/n)) = Q(cos(2π/n))(sin(2π/n))
and
[Q(cos(2π/n))(sin(2π/n)) : Q(cos(2π/n))] = 2,
with
minpoly
Q(cos(2π/n)),sin(2π/n)
(X) = X
2
+ cos
2
(2π/n) −1.
If 4  n, we can write n = 4ℓ. Then i = ζ
ℓ
n
, so i ∈ Q(ζ
n
), whence
sin(π/2ℓ) = sin(2π/n) =
ζ
n
−ζ
−1
n
i
∈ Q(ζ
n
).
Clearly
sin(π/2ℓ) ∈ Q(ζ
n
) ∩ R = Q(cos(2π/n)).
Consider the automorphism σ ∈ Gal(Q(ζ
n
)/Q)) for which σ(ζ
n
) = ζ
2ℓ+1
n
= −ζ
n
; it is easy to
see that σ has order 2. Then
σ(cos(2π/n)) = σ(cos(π/2ℓ)) = −cos(π/2ℓ),
σ(cos(π/ℓ)) = cos(π/ℓ),
σ(sin(2π/n)) = σ(sin(π/2ℓ)) =
−ζ
n
+ζ
−1
n
2(−ζ
n
)
ℓ
=
sin(π/2ℓ) if ℓ is odd,
−sin(π/2ℓ) if ℓ is even.
From this we ﬁnd that when ℓ is odd,
Q(cos(2π/n)) = Q(cos(π/2ℓ)) = Q(cos(π/ℓ))(sin(π/2ℓ)) = Q(sin(π/2ℓ)),
since cos(π/ℓ) = 1 −2 sin
2
(π/2ℓ) ∈ Q(sin(π/2ℓ)). Thus we have [Q(sin(π/2ℓ)) : Q] = 2φ(ℓ) and
Gal(Q(sin(π/2ℓ))/Q) = Gal(Q(cos(π/2ℓ))/Q) = (Z/4ℓ)
×
/{1, −1}.
Similarly, if ℓ is even,
[Q(cos(π/ℓ))(sin(π/2ℓ)) : Q(cos(π/ℓ))] = 2
and we must have
Q(cos(2π/n)) = Q(cos(π/2ℓ)) = Q(sin(π/2ℓ))
with
Gal(Q(sin(π/2ℓ))/Q) = Gal(Q(cos(π/2ℓ))/Q) = (Z/4ℓ)
×
/{1, −1}
(b) We have
sin
2
(π/12) =
1 −cos(π/6)
2
=
2 −
√
3
4
,
121
and so
sin(π/12) =
√
2 −
√
3
2
=
√
6 −
√
2
4
.
Then
Q(sin(π/12)) = Q(
√
6 −
√
2) = Q(
√
2,
√
3).
and
Gal(Q(sin(π/12))/Q))
∼
= (Z/4ℓ)
×
/{1, −1}
∼
= Z/2 ×Z/2.
Here the eﬀect of the coset of the residue class of r ∈ (Z/4ℓ)
×
is given by
r · sin(π/12) =
ζ
r
24
−ζ
r
24
i
r
= sin(rπ/12) i
1−r
.
Explicitly we have
1 · sin(π/12) = −1 · sin(π/12) = sin(π/12) =
√
6 −
√
2
4
,
5 · sin(π/12) = −5 · sin(π/12) = sin(5π/12) =
√
6 +
√
2
4
,
7 · sin(π/12) = −7 · sin(π/12) = −sin(7π/12) =
−
√
6 −
√
2
4
,
11 · sin(π/12) = −11 · sin(π/12) = −sin(11π/12) =
−
√
6 +
√
2
4
.
In terms of the generators
√
2 and
√
3 these act by
1 ·
√
2 =
√
2, 1 ·
√
3 =
√
3, 5 ·
√
2 = −
√
2, 5 ·
√
3 =
√
3,
7 ·
√
2 =
√
2, 5 ·
√
3 = −
√
3, 11 ·
√
2 = −
√
2, 11 ·
√
3 = −
√
3.
6.6. (a) We have
 Gal(Q(ζ
5
)/Q) = [Q(ζ
5
) : Q] = deg Φ
5
(X) = φ(5) = 4,
and (Z/5)
×
is cyclic generated by the residue class 2. The action is given by
2 · ζ
5
= ζ
2
5
, 2
2
· ζ
5
= ζ
4
5
, 2
3
· ζ
5
= ζ
3
5
, 2
4
· ζ
5
= ζ
5
.
(b) We have ζ
5
+ζ
−1
5
= 2 cos(2π/5) and Φ
5
(ζ
5
) = 0, so since ζ
3
5
= ζ
−2
5
and ζ
4
5
= ζ
−1
5
,
(ζ
2
5
+ζ
−2
5
) + (ζ
5
+ζ
−1
5
) + 1 = 0
and therefore
(ζ
5
+ζ
−1
5
)
2
+ (ζ
5
+ζ
−1
5
) −1 = 0.
Hence
4 cos
2
(2π/5) + 2 cos(2π/5) −1 = 0.
The quadratic polynomial 4X
2
+ 2X − 1 ∈ Z[X] has discriminant 20 which is not a square in
Q, so this is this polynomial is irreducible over Q, therefore
minpoly
Q,cos(2π/5)
(X) = X
2
+
1
2
X −
1
4
.
122
The roots of this are
−1 ±
√
5
4
. As cos(2π/5) > 0 we must have cos(2π/5) =
−1 +
√
5
4
. We
also have cos(4π/5) =
−1 −
√
5
4
. As sin(2π/5) > 0,
sin
2
(2π/5) = 1 −cos
2
(2π/5) = 1 −
1 + 5 −2
√
5
16
=
5 +
√
5
8
,
hence sin(2π/5) =
√
5 +
√
5
8
.
(c) Gal(Q(ζ
5
)/Q)
∼
= Z/4 and has 3 subgroups {1} {1, 4} Gal(Q(ζ
5
)/Q), giving the following
tower of subﬁelds.
Q(ζ
5
)
2
Q(ζ
5
)
⟨4⟩
= Q(cos(2π/5)) = Q(
√
5)
2
Q
6.7. (a) We have
ξ
2
=
(p−1)/2
∏
r=1
(ζ
r
p
−ζ
−r
p
)
2
= (−1)
(p−1)/2
(p−1)/2
∏
r=1
(ζ
r
p
−ζ
−r
p
)(ζ
−r
p
−ζ
r
p
)
= (−1)
(p−1)/2
(p−1)/2
∏
r=1
(1 −ζ
−2r
p
)(1 −ζ
2r
p
)
= (−1)
(p−1)/2
p−1
∏
r=1
(1 −ζ
2r
p
)
= (−1)
(p−1)/2
(p−1)
∏
s=1
(1 −ζ
s
p
)
since each congruence 2x ≡ t (mod p) has exactly one solution modulo p for each t.
(b) Since
(−1)
(p−1)/2
=
−1 if p ≡ 1 (mod 4),
1 if p ≡ 3 (mod 4),
and
p−1
∏
s=1
(1 −ζ
s
p
) = Φ
p
(1) = p,
the result follows.
(c) Taking square roots we ﬁnd that
ξ =
±
√
p if p ≡ 1 (mod 4),
±
√
p i if p ≡ 3 (mod 4).
As ξ ∈ Q(ζ
p
), we see that
√
p ∈ Q(ζ
p
) if p ≡ 1 (mod 4) and
√
p i ∈ Q(ζ
p
) if p ≡ 3 (mod 4).
123
6.8. Recall the wellknown formula
σ(i
1
· · · i
r
)σ
−1
= (σ(i
1
) · · · σ(i
r
)).
Then for 1 r n −2 we have
(1 2 · · · n)
r
(1 2)(1 2 · · · n)
n−r
= (1 2 · · · n)
r
(1 2)((1 2 · · · n)
r
)
−1
= (r + 1 r + 2),
while
(1 2 · · · n)
n−1
(1 2)((1 2 · · · n)
n−1
)
−1
= (1 2 · · · n)
−1
(1 2)((1 2 · · · n)
−1
)
−1
= (n 1) = (1 n).
This means that every such 2cycle (r + 1 r + 2) is in H. Also recall that every permutation
ρ ∈ S
n
is a product of 2cycles, so it suﬃces to show that every 2cycle (a b) ∈ S
n
is a product
of 2cycles of the form (r + 1 r + 2). Assuming that a < b, we also have
(a b) = (b −1 b) · · · (a + 2 a + 3)(a + 1 a + 2)(a a + 1)(a + 1 a + 2)(a + 2 a + 3) · · · (b −1 b),
and this is in H. Hence H = S
n
.
6.9. (a) For each u ∈ E,
σ(T(u)) = σ(u +σ(u) +σ
2
(u) +· · · +σ
n−1
(u))
= σ(u) +σ
2
(u) +· · · +σ
n
(u)
= σ(u) +σ
2
(u) +· · · +σ
n−1
(u) +u = T(u),
so T(u) is ﬁxed by σ and all its powers, hence by Gal(E/K). Therefore T(u) is in E
Gal(E/K)
= K.
It is straightforward to verify that the resulting function Tr
E/K
: E −→ K is Klinear.
(b) Let v ∈ E and suppose that Tr
E/K
(v) = 0. By Artin’s Theorem 6.15, the linear combination
of characters id +σ +· · · +σ
n−1
must be linearly independent, so there is an element t ∈ E for
which
Tr
E/K
t = t +σ(t) +· · · +σ
n−1
(t) ̸= 0.
Then
u = vσ(t) + (v +σ(v))σ
2
(t) +· · · + (v +σ(v)σ
2
(t) +· · · +σ
n−2
(v))σ
n−1
(t)
satisﬁes
u −σ(u) = v
σ(t) +σ
2
(t) +· · · +σ
n−1
(t)
−
σ(v) +· · · +σ
n−1
(v)
t
= v
t +σ(t) +σ
2
(t) +· · · +σ
n−1
(t)
−
v +σ(v) +· · · +σ
n−1
(v)
t
= (Tr
E/K
t)v −(Tr
E/K
v)t = (Tr
E/K
t)v.
So we obtain
v =
1
Tr
E/K
t
u −σ
1
Tr
E/K
t
u
.
6.10. (a) This can be proved by induction on n. Write
e
[m]
r
=
∑
i
1
<i
2
<···<i
r
m
X
i
1
· · · X
i
r
, s
[m]
r
=
∑
1im
X
r
i
.
Then we easily ﬁnd that
e
[m]
r
= e
[m−1]
r
+e
[m−1]
r−1
X
m
, s
[m]
r
= s
[m−1]
r
+X
r
m
.
124
Notice also that e
[m]
r
= 0 whenever r > m. The desired result is that for all n 1 and k 1,
s
[n]
k
= e
[n]
1
s
[n]
k−1
−e
[n]
2
s
[n]
k−2
+· · · + (−1)
k−1
e
[n]
k−1
s
[n]
1
+ (−1)
k
ke
[n]
k
.
When n = 1 we have s
[1]
r
= X
r
1
and e
[1]
1
= X
1
from which the result follows. Now suppose that
the result is true for some n 1. Then s
[n+1]
k
= s
[n]
k
+X
k
n+1
, while
e
[n+1]
1
s
[n+1]
k−1
−e
[n+1]
2
s
[n+1]
k−2
+· · · + (−1)
k−1
e
[n+1]
k−1
s
[n+1]
1
+ (−1)
k
ke
[n+1]
k
=
(e
[n]
1
+X
n+1
)(s
[n]
k−1
+X
k−1
n+1
) −(e
[n]
2
+e
[n]
1
X
n+1
)(s
[n]
k−2
+X
k−2
n+1
) +· · ·
+ (−1)
k−1
(e
[n]
k−1
+e
[n]
k−2
X
n+1
)(s
[n]
1
+X
n+1
) + (−1)
k
k(e
[n]
k
+e
[n]
k−1
X
n+1
)
= s
[n]
k
+ (e
[n]
1
X
k−1
n+1
−e
[n]
2
X
k−2
n+1
+· · · + (−1)
k−1
e
[n]
k−1
X
n+1
)
+ (s
[n]
k−1
−e
[n]
1
s
[n]
k−2
+· · · + (−1)
k−1
e
[n]
k−2
s
[n]
1
+ (−1)
k
ke
[n]
k−1
)X
n+1
+ (X
k
n+1
−e
[n]
1
X
k−1
n+1
+· · · + (−1)
k−1
e
[n]
k−2
X
2
n+1
)
= s
[n]
k
+X
k
n+1
= s
[n+1]
k
,
which demonstrates the inductive step.
(b)(i) We have h
1
= e
1
, h
2
= e
2
1
−e
2
and h
3
= e
3
−2e
1
e
2
+e
3
1
.
(ii) This can be done by induction on n in a similar way to part (a).
125
√ Q( 3 2)
N
ggg n ggggg nnn ggggg nnnnn ggg 2 ggggg nnn 2 √ ggggg 3
√ Q( 3 2, ζ3 )
H
3 yyy N ee ee yyy ee yyy ee yyy ee 3 yyy 3 ee yyy e yyy yyy eee yyy ee yyy ee e yyy
Q( 2 ζ )
pp pp pp 2 pp pp √ 2 pp3 pp Q( 3 2 ζ3 ) pp I pp pp pp p
3
Q
S
kk kkkk 2 kkkk kk kkkk kkkk
Q(ζ3 )
Q
w ww ww w ww 3 www w {id, (2 3)} {id, (1 2)} {id, (1 3)} ww ww ww ww 2 2 2 ww w
ooo }} ooo }}} oo ooo3 }}} 3 oooo }} o 3 }} ooo } ooo }} oo } }} ooo }} ooo
Gal(E/Q) ∼ S3 =
2
{id, (1 2 3), (1 3 2)}
{id}
√ The Galois Correspondence for Q( 3 2, ζ3 )/Q
Introduction: What is Galois Theory? Much of early algebra centred around the search for explicit formulae for roots of polynomial equations in one or more unknowns. The solution of linear and quadratic equations in a single unknown was well understood in antiquity, while formulae for the roots of general real cubics and quartics was solved by the 16th century. These solutions involved complex numbers rather than just real numbers. By the early 19th century no general solution of a general polynomial equation ‘by radicals’ (i.e., by repeatedly taking nth roots for various n) was found despite considerable eﬀort by many outstanding mathematicians. Eventually, the work of Abel and Galois led to a satisfactory framework for fully understanding this problem and the realization that the general polynomial equation of degree at least 5 could not always be solved by radicals. At a more profound level, the algebraic structure of Galois extensions is mirrored in the subgroups of their Galois groups, which allows the application of group theoretic ideas to the study of ﬁelds. This Galois Correspondence is a powerful idea which can be generalized to apply to such diverse topics as ring theory, algebraic number theory, algebraic geometry, diﬀerential equations and algebraic topology. Because of this, Galois theory in its many manifestations is a central topic in modern mathematics. In this course we will focus on the following topics. • The solution of polynomial equations over a ﬁeld, including relationships between roots, methods of solutions and location of roots. • The structure of ﬁnite and algebraic extensions of ﬁelds and their automorphisms. We will study these in detail, building up a theory of algebraic extensions of ﬁelds and their automorphism groups and applying it to solve questions about roots of polynomial equations. The techniques we will meet can also be applied to study the following some of which may be met by people studying more advanced courses. • Classic topics such as squaring the circle, duplication of the cube, constructible numbers and constructible polygons. • Applications of Galois theoretic ideas in Number Theory, the study of diﬀerential equations and Algebraic Geometry. There are many good introductory books on Galois Theory, some of which are listed in the Bibliography. In particular, [2, 3, 8] are all excellent sources and have many similarities to the present approach to the material.
c ⃝A. J. Baker (2009)
ii
Finding roots of complex polynomials of small degree 1. Identifying irreducible polynomials 1. Fields and their extensions 2.3.1. Polynomial rings 1. Galois groups of even and odd permutations 4. convention.4. Algebraic extensions 3. Monomorphisms between extensions 3. Galois extensions inside the complex numbers and complex conjugation 4.5. Multiplicity of roots and separability 3.2.5. Recollections on integral domains and ﬁelds 1.4.2. Splitting ﬁelds and Kronecker’s Theorem 3. Subﬁelds of Galois extensions and relative Galois groups 4. Integral domains.1.5.7.6. Normal extensions and splitting ﬁelds Exercises on Chapter 3 Chapter 4. Finite ﬁelds
iii
ii 1 1 1 6 12 16 19 23 27 27 29 33 35 35 39 42 45 48 52 54 55 57 57 58 60 61 62 64 65 68 71 73 73
. Working with Galois groups 4. Simple and ﬁnitely generated extensions Exercises on Chapter 2 Chapter 3. Galois extensions 4.1.1. The Primitive Element Theorem 3. Fields and subﬁelds 2.8. ﬁelds and polynomial rings Basic notions.6.Contents
Introduction: What is Galois Theory? Chapter 1. Automorphisms of rings and ﬁelds Exercises on Chapter 1 Chapter 2. Galois extensions and the Galois Correspondence 4. Galois extensions for ﬁelds of positive characteristic 5.1. Algebraic closures 3. Subgroups of Galois groups and their ﬁxed ﬁelds 4. etc 1. Kaplansky’s Theorem Exercises on Chapter 4 Chapter 5.2. Algebraic extensions of ﬁelds 3.2. The Galois Correspondence and the Main Theorem of Galois Theory 4.4.7.3.3.
Artin’s Theorem on linear independence of characters 6. Solvability and radical extensions 6. A proof of the Fundamental Theorem of Algebra 6. Symmetric functions Exercises on Chapter 6 Bibliography Solutions Chapter Chapter Chapter Chapter Chapter Chapter
77 79 80 83 83 84 88 90 92 96 97 101 103 103 110 112 114 117 119
1 2 3 4 5 6
iv
. The trace and norm mappings Exercises on Chapter 5 Chapter 6.6.3.1.5. A Galois Miscellany 6.2.3. Cyclotomic extensions 6.5.2. Simple radical extensions 6.4. Galois groups of ﬁnite ﬁelds and Frobenius mappings 5.
R[X] and C[X] are all integral domains.e. (i) (ii) (iii) (iv) The ring of integers. An ideal I ▹ R will always mean a twosided ideal.e. If p is a prime. if for every s ∈ S there is an r ∈ R with φ(r) = s. Details may be found in [3. if for r1 . i. 1. A commutative ring R in which there are no zerodivisors is called an integral domain or an entire ring. where R is an integral domain.. and complex numbers. i. ﬁelds and polynomial rings
Basic notions.
1
. • φ is an epimorphism if it is surjective. a ring will always be a ring with unity 1 ̸= 0.. uv ∈ I =⇒ u ∈ I or v ∈ I.3. Recollections on integral domains and ﬁelds The material in this section is standard and most of it should be familiar. Fp = Z/p = Z/(p). Under a ring homomorphism φ : R −→ S. Definition. 5] or other books containing introductory ring theory.CHAPTER 1
Integral domains. convention. Z. the ring of integers modulo p. φ(1) = 1. 1. • φ is a monomorphism if it is injective. i.. v ∈ R. This means that for u. φ(r1 ) = φ(r2 ) =⇒ r1 = r2 . The rings of rational numbers. or equivalently if I R. in particular. Example. i. The polynomial ring R[X]. Definition. Let I ▹ R be a proper ideal in a commutative ring R. 1.e. etc In these notes.
• I is a maximal ideal R if whenever J ▹ R is a proper ideal and I ⊆ J then J = I.
1..
or equivalently if ker φ = {0}.1. the polynomial rings Z[X].e. real numbers. Most of the rings encountered will also be commutative.2. Q. Let φ : R −→ S be a ring homomorphism. • I is a prime ideal if for u.4. uv = 0 =⇒ u = 0 or v = 0. The following rings are integral domains.
1. An ideal I ▹ R in a ring R is proper if I ̸= R. R.1. r2 ∈ R. Definition. C. Q[X]. if it is invertible (in which case its inverse is also an isomorphism). 1 ∈ R is sent to 1 ∈ S. • φ is an isomorphism if it is both a monomorphism and an epimorphism. v ∈ R.
The quotient monomorphism η : Z/ ker η −→ R identiﬁes Z/ ker η with the subring im η = im η of the integral domain R. −n 0 if n = 0.. 1.8. the rings Z and Fp = Z/p = Z/(p) for p > 0 a prime are called prime rings. this p is called the characteristic of R and denoted char R. its characteristic char R is a prime. But every subring of an integral domain is itself an integral domain. there is a unique nonnegative integer p 0 such that ker η = (p). Lemma.6(i).5.9. we can consider it as containing as a subring of the form Z/(char R) since the quotient homomorphism η : Z/(char R) −→ R gives an isomorphism Z/(char R) −→ im η.6. Hence Z is a principal ideal domain. p is a prime. If n n is a prime. Proposition. Since 1 ∈ R is nonzero. Now by Proposition 1. n 0. Example. So suppose that p > 0. By Example 1. When discussing a ring with unit R. Consider p = char R. (ii) The quotient ring R/I is a ﬁeld if and only if I is a maximal ideal. so I = (n) for some n ∈ Z which we can always take to be nonnegative. In particular.7. 1. i. the quotient ring Z/n = Z/(n) is an integral domain if and only if
For any (not necessarily commutative) ring with unity there is an important ring homomorphism η : Z −→ R called the unit or characteristic homomorphism which is deﬁned by 1 + · · · + 1 if n > 0. The rings Z and Z/n = Z/(n) for n > 0 are often called core rings. then (q) = (p) if and only if q = up for some unit u ∈ R. this is equivalent to the requirement that whenever p  xy with x. (i) The quotient ring R/I is an integral domain if and only if I is a prime ideal. Example. q ∈ R.5.e. 1. When considering integral domains. allowing us to identify these rings. hence Z/ ker η is an integral domain. every integral domain contains as a subring either Z = Z/(0) (if char R = 0) or Z/(p) if p = char R > 0 is a nonzero prime. 1.
2
. ker η ▹ Z is a proper ideal and using the Isomorphism Theorems we see that there is a quotient monomorphism η : Z/ ker η −→ R which allows us to identify the quotient ring Z/ ker η with the image ηZ ⊆ R as a subring of R. y ∈ R then p  x or p  y. Notice that if p. • p ∈ R is prime if (p) ▹ R is a prime ideal. If p = 0 we are done. n η(n) = n1 = −(1 + · · · + 1) if n < 0. This subring is sometimes called the characteristic subring of R. We also write p  x if x ∈ (p).• I ▹ R is principal if I = (p) = {rp : r ∈ R} for some p ∈ R. • R is a principal ideal domain if it is an integral domain and every ideal I ▹R is principal. Let R be a commutative ring and I ▹ R an ideal. 0.7. ker η = (p) is prime ideal and so by Example 1. Every ideal I ▹Z is principal. 1. Proof. If R is an integral domain. Remark.
hence pt = 0 for any t ∈ R. The familiar rings Q.1) (u + v) = u + u v+ u v + ··· + uv p−1 + v p . j j! (p − j)! j! (p − j)!
There are no factors of p appearing in (p − 1)!. Definition.2b) Hence in R we have ( ) p ≡ 0 (mod p). This is equivalent to requiring that k× = k − {0}.13. We have p1 = 0 in R. there is and element v ∈ R for which uv = 1 = vu. A commutative ring k is a ﬁeld if every nonzero element u ∈ k is a unit.12.
1.11. we obtain the required equation in R.14.. so since this number is an integer it must be divisible by p.2) with the expansion of (1.Here is a useful and important fact about rings which contain a ﬁnite prime ring Fp .2a) p . R and C are all ﬁelds. Theorem (Idiot’s Binomial Theorem). ( ) p (1. j! or (p − j)!.. Let R be a ring. 1. If u ̸= 0. j
( ) p 1 = 0. Let k be a ﬁeld. Let R be a commutative ring with unit containing Fp for some prime p > 0. If n 1. 1. Then we have ( ) p p (p − 1)! (p − 1)! = =p× . An element u ∈ R is a unit if it is invertible. So at least one of u. We usually write u−1 for this element v.
3
. Proof. Suppose that u.1). We will denote the set of all invertible elements of R by R× and note that it always forms a group under multiplication. then (u + v)p = up + v p . 1. v must be 0. which is necessarily unique and is called the (multiplicative) inverse of u in R. the quotient ring Z/n is a ﬁeld if and only if n is a prime.e. v ∈ k and uv = 0.e. The Binomial Expansion yields ( ) ( ) ( ) p p−1 p p−2 2 p p p (1. 1 2 p−1 Now suppose that 1 j p − 1. v ∈ R. i. hence v = 0. Proposition. 1. j or equivalently (1. Definition. Example. Every ﬁeld is an integral domain. j Combining the divisibility conditions of (1. i. (u + v)p = up + v p . If u. we can multiply by to obtain v = u−1 uv = 0.10. u−1 Proof.
(i) There is a ﬁeld of fractions of D. (i) Consider the set P(D) = {(a. showing that v is a unit with inverse t. 1. bd].
1. there is a unique homomorphism φ : Fr(D) −→ F such that φ(t) = φ(t) for all t ∈ D ⊆ Fr(D). bd). b] + [c. bd). b][c. This is certainly true for the integers Z which are contained in the ﬁeld of rational numbers Q. there is always a ‘smallest’ such ﬁeld which is unique up to an isomorphism. b] and the set of equivalence classes by Fr(D). Then p  u or p  v. then (a′ d′ + b′ c′ )bd = a′ d′ bd + b′ c′ bd = ab′ d′ d + b′ bcd′ = (ad + bc)b′ d′ . namely (a′ . d] = [ad + bc. Theorem. the prime ring Fp is itself a ﬁeld. A similar calculation shows that (a′ c′ . Now let D be an integral domain. b) ⇐⇒ ab′ = a′ b. b) : a. The ﬁelds Q and Fp where p > 0 is a prime are the prime ﬁelds. b′ d′ ) ∼ (ad + bc. b′ ] = [a. this is left as an exercise. b′ d′ ) ∼ (ac. Now introduce an equivalence relation ∼ on P(D). Proof. so multiplication is also well deﬁned. if [a′ . 1. [a. D inc
φ
/F <
∃! φ
Fr(D) Proof. and so (a′ d′ + b′ c′ . Definition. d]. Fr(D). v ∈ R. v ∈ R. (ii) If φ : D −→ F is a ring monomorphism into a ﬁeld F .15. which contains D as a subring. For example.16. p = uv =⇒ u or v is a unit. d′ ] = [c. b] and [c′ . Let D be an integral domain.
Of course.A nonzero element p ∈ R is irreducible if for u. it is necessary to check that this relation is an equivalence relation. b ̸= 0}. and for a prime p > 0. We deﬁne addition and multiplication on Fr(D) by [a.17. We denote the equivalence class of (a. bd]. Then (1 − tv)p = 0 and so tv = 1. b) by [a. Let R be an integral domain. It is now straightforward
4
. Of course. b ∈ D. A natural question to ask is whether D is isomorphic to a subring of a ﬁeld. If p ∈ R is a nonzero prime then it is an irreducible. hence addition is well deﬁned. and we might as well assume that u = tp for some t ∈ R. d] = [ac. we can view Z as a subring of any subﬁeld of the complex numbers so an answer to this question may not be unique! However. b′ ) ∼ (a. Suppose that p = uv for some u. Lemma.
We need to verify that these operations are well deﬁned.
We can view D as a subring of Fr(D) using the map j : D −→ Fr(D). b′ ) = φ(a′ )φ(b′ )−1 = φ(a′ )φ(b)φ(b)−1 φ(b′ )−1 = φ(a′ b)φ(b)−1 φ(b′ )−1 = φ(ab′ )φ(b′ )−1 φ(b)−1 = φ(a)φ(b′ )φ(b′ )−1 φ(b)−1 = φ(a)φ(b)−1 = Φ(a. 1. 1] ∈ Fr(D) and D with the subring im j ⊆ Fr(D). b] = [0. 1. Let [a. [b. 1]. If D is a subring of a ﬁeld F . [0. Now let [a. The next three corollaries are left as an exercise. Then b ̸= 0 and [a. We also have [a. Then [a. notice that for any b ̸= 0. as we will soon see. Φ(a. then Fr(D) ⊆ Fr(F ) = F and Fr(D) is the smallest subﬁeld of F containing D. 1] = 1.to show that Fr(D) is a commutative ring with zero 0 = [0. Corollary. b] ∈ Fr(D). it is easy to check that it is a monomorphism.18. 1]. a ̸= 0. (ii) Consider the function Φ : P(D) −→ F . b). b] = [0. Then there is a unique induced ring homomorphism φ∗ : Fr(D1 ) −→ Fr(D2 ) which satisﬁes φ∗ (t) = φ(t) whenever t ∈ D1 ⊆ Fr(D1 ). b] has [b.. b).
It is now easy to verify that φ is a ring homomorphism which agrees with φ on D ⊆ Fr(D). In fact. D1
φ
/ D2
inc φ∗
inc
Fr(D1 )
/ Fr(D2 )
Moreover. This shows that Fr(D) is a ﬁeld.
5
. b]. Therefore we may identify t ∈ D with j(t) = [t. Let D1 and D2 be integral domains and let φ : D1 −→ D2 be a ring monomorphism. b][b. a] as an inverse. this construction has the following properties.19. b] ∈ Fr(D) be nonzero. Corollary. so [a. b) which is equivalent to requiring that a = 0. If F is a ﬁeld then F = Fr(F ). If (a′ . Hence we may deﬁne the function φ : Fr(D) −→ F . 1]
which is a ring homomorphism. 1] if and only if (0. 1.20. i. 1] if and only if a = b. a] ∈ Fr(D) satisfy [a. 1] and unit 1 = [1. Fr(D) is a ﬁeld. b) = φ(a)φ(b)−1 . 1) ∼ (a. ba] = [1.e. φ([a. so Φ is constant on each equivalence class of ∼. b). b]) = Φ(a. b′ ) ∼ (a. b] = [1. then Φ(a′ . j(t) = [t. a] = [ab. Corollary.
. R[X1 . . a1 a2 a1 b2 + a2 b1 + = . . We identify r ∈ R with the obvious constant polynomial. b).. such p(X) are called polynomials. p1 . Remarks. Xn−1 ][Xn ]
6
.20 is sometimes said to imply that the construction of Fr(D) is functorial in the integral domain D. Polynomial rings Let R be a commutative ring. the identity homomorphism id : D −→ D induces the identity homomorphism (id)∗ = id : Fr(D) −→ Fr(D). . .• If φ : D1 −→ D2 and θ : D2 −→ D3 are monomorphisms between integral domains then θ∗ ◦ φ∗ = (θ ◦ φ)∗ as homomorphisms Fr(D1 ) −→ Fr(D3 ). b1 b2 b1 b2 a1 a2 a1 a2 a1 a2 × = = . More generally. (a) When working with a ﬁeld of fractions we usually adopt the familiar notation a = a/b = [a. Then R[X] is a commutative ring with the constant polynomials 0 and 1 as its zero and unit. Xn over R. pm ∈ R. however as the word quotient is also associated with quotient rings we prefer to avoid using that terminology. • For any integral domain D. . . b] b for the equivalence class of (a. . . . b1 b2 b1 b2 b1 b2
The ﬁeld of fractions of an integral domain is sometimes called its ﬁeld of quotients. Addition and multiplication in R[X] are deﬁned by (p0 + p1 X + · · · + pm X m ) + (q0 + q1 X + · · · + qm X m ) = (p0 + q0 ) + (p1 + q1 )X + · · · + (pm + qm )X m . . 1. . this allows us to view R as a subring of R[X] and the inclusion function inc : R −→ R[X] is a monomorphism. φ
/ D2
D1
θ
/ D3
D
id
/D
inc φ∗
inc θ∗
inc
inc
inc
Fr(D1 )
/ Fr(D2 )
/ Fr(D3 )
/ Fr(D) Fr(D) id∗ = id
1. We will make frequent use of the ring R[X] of polynomials over R in an indeterminate X. .21. Xn ] = R[X1 .2. . The rules for algebraic manipulation of such symbols are the usual ones for working with fractions. This consists of elements of form p(X) = p0 + p1 X + · · · + pm X m where m 0 and p0 . . i. (b) Corollary 1. we inductively can deﬁne the ring of polynomials in n indeterminates X1 . and (p0 + p1 X + · · · + pm X m )(q0 + q1 X + · · · + qm X m ) = (p0 q0 ) + (p0 q1 + p1 q0 )X + · · · + (p0 qm + p1 qm−1 + · · · + pm−1 q1 + pm q0 )X 2m . .e.
.
It is then straightforward to check that φs is a ring homomorphism with the stated properties and moreover is the unique such homomorphism. then εs1 .. (i) For each s ∈ S there is a unique ring homomorphism φs : R[X] −→ S for which • φs (r) = φ(r) for all r ∈ R. . .sn : R[X1 .. . .. . sn ∈ S.. . . ..... . . . . . . Xn ] which sends each element of R to itself considered as a constant polynomial.sn R[X1 . .. . Xn ] −→ S for which • φs1 .. . . .. .. sn . sn ∈ S..sn = incs1 .. there is a unique ring homomorphism φs1 .. . . Xn ] −→ R. . . . .. . 1. R inc
φ
/S =
∃! φs
R[X] (ii) For n 1 and s1 .. Theorem (Homomorphism Extension Property). . Xn ] Proof.for n 1.. .sn (Xi ) = si for i = 1. (Sketch) (i) For a polynomial p(X) = p0 + p1 X + · · · + pm X m ∈ R[X]. Again there is an inclusion monomorphism inc : R −→ R[X1 . Xn ] ⊆ S... sn and we denote its image by R[s1 . .22. Xn ] −→ S is called evaluation at s1 .. . . .sn as the extension of φ by evaluation at s1 . .. rn ∈ R... . .3) φs (p(X)) = p0 + p1 s + · · · + pm sm ∈ S. It is standard to write p(s1 . ... .. Slightly more generally we may take the inclusion of a subring inc : R −→ S and s1 . sn ] = εs1 .. . • φs1 . .sn
R[X1 . Let φ : R −→ S be a ring homomorphism. . . . ..sn (p(X1 . These polynomial rings have an important universal property. then we have the homomorphism εr1 . (ii) is proved by induction on n using (i). we deﬁne (1. .. . .. . . .
7
.. ...sn (r) = φ(r) for all r ∈ R... .. We will refer to φs1 . R inc
φ
8/ S
∃! φs1 .. n. An extremely important special case occurs when we start with the identity homomorphism id : R −→ R and r1 . .. . • φs (X) = s.rn = idr1 . .. .. sn ) = φs1 . Xn )). ..sn : R[X1 . .. ..rn : R[X1 .
. .Then R[s1 .
1. Here is an example illustrating how we will use such evaluation homomorphisms. . Xn ] of polynomials in the indeterminates X1 . called the subring generated by s1 . . Then there are unique polynomials q(X). . As we will make considerable use of such rings we describe in detail some of their important properties. 1. . hence we also have Q[i] ∼ Q[X]/(X 2 + 1). Notice also that in these examples we have ker ε−i = ker εi = (X 2 + 1) ▹ Q[X]. In the situation discussed in this result.25. These evaluation homomorphisms are related by complex conjugation since ε−i (p(X)) = εi (p(X)).23. 1. Also. Proposition. later we will write Q(i) for this subﬁeld. = In fact (X 2 + 1) is actually a maximal ideal and so Q[i] ⊆ C is a subﬁeld. For k = Q. First we recall long division in a polynomial ring k[X] over a ﬁeld k. . ε−i . . Xn ] in the indeterminates X1 . We easily see that εi Q[X] ⊆ C is a subring Q[i] ⊆ C consisting of the complex numbers of form a + bi with a. . we would also have ε−i Q[X] = Q[i]. . r(X) ∈ k[X] for which f (X) = q(X)d(X) + r(X) and either deg r(X) < deg d(X) or r(X) = 0. Example. 1. . Consider the inclusion homomorphism inc : Q −→ C. Then the polynomial ring k[X1 . sn ] is a subring of S. b ∈ Q. (i) The ring R[X] of polynomials in an indeterminate X over R is an integral domain. . Xn over R is an integral domain. .
8
. . . which is equivalent to the functional equation ε−i = ( ) ◦ εi . We have the evaluation at i homomorphism εi . . evaluation at −i. Let k be a ﬁeld. Example. Xn is an integral domain. Let k be a ﬁeld and n 1. d(X) ∈ k[X] and assume that d(X) ̸= 0 so that deg d(X) > 0. Theorem (Long Division). Notice that if we had used −i instead of i. d(X) = the divisor . sn over R. for which εi (X) = i. Corollary. the following names are often used. (ii) The ring R[X1 . ﬁnd the quotient and remainder when f (X) = 6X 4 − 6X 3 + 3X 2 − 3X + 1 is divided by d(X) = 2X 2 + 1. 1. . .24. . Let f (X). full details can be found in a basic course on commutative rings or any introductory book on this subject. r(X) = the remainder . . .27. . . f (X) = the dividend . .26. q(X) = the quotient. We refer to the process of ﬁnding q(X) and r(X) as long division of f (X) by d(X). Let R be an integral domain.
g(X) is the monic polynomial of greatest degree which divides both of f (X).29. 1.Solution. 3X 2 + 2X 2X 2 + 1 6X 4 + 4X 3 + 3X 2 + 2X + 1 6X 4 + 0X 3 + 3X 2 + 0X + 0 4X 3 + 0X 2 + 2X + 1 4X 3 + 0X 2 + 2X + 0 1 Hence 6X 4 − 6X 3 + 3X 2 − 3X + 1 ≡ (3X 2 + 2X)(2X 2 + 1) + 1 (mod 5). First notice that working modulo 5 we have f (X) = 10X 5 + 6X 4 − 6X 3 + 3X 2 − 3X + 1 ≡ X 4 + 4X 3 + 3X 2 + 2X + 1 (mod 5). giving q(X) = 3X 2 + 2X and r(X) = 1. Notice also following multiplicative inverses in F5 : 2−1 ≡ 3 (mod 5). ﬁnd the quotient and remainder when f (X) = 10X 5 + 6X 4 − 6X 3 + 3X 2 − 3X + 1 is divided by d(X) = 2X 2 + 1. 1. b(X) ∈ k[X] such that a(X)f (X) + b(X)g(X) = gcd(f (X). Here the greatest common divisor gcd(f (X). Example. Solution. Corollary. g(X) ∈ k[X] be nonzero. giving q(X) = 3X 2 − 3X and r(X) = 1.
9
3−1 ≡ 2 (mod 5).
4−1 ≡ 4 (mod 5).
. Let f (X). Let k be a ﬁeld and X an indeterminate. In the usual notation we have the following calculation. Then there are a(X). An important consequence of Theorem 1. g(X)) of f (X).26 is the following which makes use of the Euclidean Algorithm. For k = F5 . 3X 2 − 3X 2X 2 + 1  6X 4 − 6X 3 + 3X 2 − 3X + 1 6X 4 + 0X 3 + 3X 2 + 0X + 0 − 6X 3 + 0X 2 − 3X + 1 − 6X 3 + 0X 2 − 3X + 0 1 Hence 6X 4 − 6X 3 + 3X 2 − 3X + 1 = (3X 2 − 3X)(2X 2 + 1) + 1. g(X). We have the following calculation.28. g(X)).
(ii) This follows from Proposition 1. Then a nonconstant polynomial p(X) ∈ k[X] is an irreducible if and only if it is a prime. 1.29 we can ﬁnd suitable a(X). Since p(X) and h(X) are in the ideal I. Suppose that p(X) is irreducible (and hence is nonzero). this would contradict the minimality of d. i. v(X) ∈ k[X]. b(X) ∈ k[X] for which / a(X)p(X) + b(X)q(X) = gcd(p(X). Then by Corollary 1. Proof.31. Let k be a ﬁeld and X an indeterminate.30. gcd(p(X). u(X)) = p(X) or gcd(p(X). (ii) The ideal (p(X)) ▹ k[X] is prime if and only if p(X) = 0 or p(X) is irreducible in k[X]. so we must have r(X) = 0. But gcd(p(X).e. This shows that in the quotient ring k[X]/(p(X)) the residue class of q(X) has the residue class of b(X) as its inverse. b(X) ∈ k[X] such that a(X)p(X) + b(X)u(X) = gcd(p(X). (i) Let I ▹ k[X] and assume that I ̸= (0).
10
. p(X) is an irreducible by (iii). (iv) Since k[X] is an integral domain and not a ﬁeld. and so p(X) is prime. Now let p(X) ∈ I. I = (h(X)) for some h(X) ∈ k[X]. Theorem. by Corollary 1.6(i). there are a(X).29. Thus I ⊆ (p(X)) ⊆ I and therefore I = (p(X)). Proof. say d = deg h(X). (iii) This follows from Proposition 1. there are q(X). Then for any q(X) ∈ k[X] with q(X) ∈ (p(X)).30. q(X)) = 1 since p(X) is irreducible. So suppose that p(X) is irreducible and that p(X)  u(X)v(X) for u(X).15 we already know that p(X) is irreducible if it is prime. By Long Division. a(X)p(X) + b(X)u(X) = 1. we also have r(X) = p(X) − q(X)h(X) ∈ I. If r(X) ̸= 0. (iii) The quotient ring k[X]/(p(X)) is an integral domain if and only if p(X) = 0 or p(X) is irreducible in k[X]. it follows that if k[X]/(p(X)) is a ﬁeld then because it is an integral domain. Then there must be at least one element of I with positive degree and so we can choose h(X) ∈ I of minimal degree.. (iv) The quotient ring k[X]/(p(X)) is a ﬁeld if and only if p(X) is an irreducible in k[X]. In the latter case. By Lemma 1. showing that p(X) = q(X)h(X). But since p(X) is irreducible. so a(X)p(X) + b(X)q(X) = 1. u(X)) = 1. (i) Every ideal I ▹ k[X] is principal. This shows that p(X)  u(X) or p(X)  v(X). Let k be a ﬁeld and X an indeterminate. Proposition. r(X) ∈ k[X] such that p(X) = q(X)h(X) + r(X) and deg r(X) < d or r(X) = 0. q(X)). u(X)). and multiplying through by v(X) gives a(X)p(X)v(X) + b(X)u(X)v(X) = v(X) and so p(X)  v(X).1.
where c ∈ k. . If deg f (X) > 1. p(X) = p0 + p1 X + · · · + pd X d = pd q(X). 1. . . Moreover. where u1 . . hence must equal it. . Corollary. mi ).
11
. . notice that if p(X) ∈ k[X]. . pk (X) is unique apart from the order of the terms. Proposition (Unique Factorization Property). . Remark. . and therefore deg fj (X) < deg f (X). c is unique and the sequence of polynomials p1 (X). . In particular. where q(X) = p−1 p0 + p−1 p1 X + · · · + p−1 pd−1 X d−1 + X d . ud ∈ k. (p(X)) = (q(X)). This gives the inductive step.30. then f (X) = c(X − v1 )m1 · · · (X − vr )mr . Corollary. If k is a ﬁeld and X an indeterminate. Then the sequence of roots u1 . then provided d = deg p(X) > 0. . qj (X) ∈ k[X] are irreducible monic polynomials. suppose that p1 (X) · · · pk (X) = q1 (X) · · · qℓ (X) where pi (X). Suppose that f (X) ∈ k[X] factors into linear factors f (X) = c(X − u1 ) · · · (X − ud ). (Sketch) Existence is proved by induction on the degree of f (X) and begins with the obvious case deg f (X) = 1.34.31(i). hence divides one of the qj (X). vr are the distinct roots. In connection with Theorem 1. Then by Proposition 1.33. The number of distinct roots of a nonconstant polynomial f (X) ∈ k[X] is at most deg f (X). 1. . . k(X). . . Definition. where mi > 0 and this factorization is unique apart from the order of the pairs (vi . Every nonconstant polynomial f (x) ∈ k[X] has a factorization f (x) = cp1 (X) · · · pk (X). . then the ﬁeld of fractions of k[X] is the ﬁeld of rational functions. d d d This easily implies that as ideals of k[X]. and p1 (X). After cancelling common factors we obtain qk+1 (X) · · · qℓ (X) = 1. we have for some pd ̸= 0. bj ∈ k and b0 + b1 X + · · · + bn X n ̸= 0. ud is unique apart from the order. The elements of k(X) are fractions of the form a0 + a1 X + · · · + am X m b0 + b1 X + · · · + bn X n with ai . and this monic polynomial is unique. . 1. if v1 . So we can always ﬁnd a monic polynomial as the generator of a given ideal. Proof.32. or f (X) = f1 (X)f2 (X) with both factors of positive degree. . . pk (X) ∈ k[X] are irreducible monic polynomials. . 1.1. To prove uniqueness. then either f (X) is already irreducible.35. By reordering we can assume that pi (X) = qi (X) and k ℓ. each pi (X) is prime.36. and so we see that k = ℓ.
• a0 ̸≡ 0 (mod p2 ). . Hence
Also. d − 1.2a). Our next result is a special case of the Eisenstein Irreducibility Test. 1. Then f (X) has a proper factorisation over Z if and only it has a proper factorisation over Q. where d = deg f (X). Let p 2 be a prime. Then f (X) is irreducible in Q[X] and hence also in Z[X]. 1
12
. Working in Z[X]. • ad ̸≡ 0 (mod p). h(X) ∈ R[X] with deg g(X) > 0 and deg h(X) > 0. If R = Z or Q. . Let f (X) ∈ Z[X]. we say that f (X) has a proper factorisation over R if f (X) = g(X)h(X) for some g(X). If f (X) ∈ Z[X] then we can also consider f (X) as an element of Q[X]. Suppose that p > 0 is a prime for which the following three conditions hold: • ak ≡ 0 (mod p) for k = 0. So to ﬁnd factors of f (X) it is suﬃcient to look for factors in Z[X]. 1. Then the polynomial Φp (X) = 1 + X + · · · + X p−1 ∈ Z[X] is irreducible in Q[X] and hence also in Z[X]. Let us consider factorisation of polynomials over Q. .1. . The version here is slightly more general than the more usual one which corresponds to taking s = 0. Identifying irreducible polynomials When k is a ﬁeld.39.
( ) p = p ̸≡ 0 (mod p2 ). Let f (X) ∈ Z[X] and s ∈ Z. Example. Proposition (Gauss’s Lemma).
( ) p p! = k! (p − k)! k Φp (X) ≡ (X − 1)p−1 (mod p)
when k = 1. Choose ai ∈ Z so that f (X) = a0 + a1 (X − s) + · · · + ad−1 (X − s)d−1 + ad (X − s)d . p divides
(mod p). 1. Φp (X)(X − 1) = (1 + X + · · · + X p−1 )(X − 1) = Xp − 1 = (1 + (X − 1))p − 1 p ∑ (p) = (X − 1)k k
k=1
≡ (X − 1)p since by (1. . we will need some eﬀective methods for deciding when a polynomial in k[X] is irreducible. . Proposition (Eisenstein Test). .3. Proof. p − 1.37. .38.
5a) Xn − 1 = Φd (X). X 2 − 1 = (X − 1)(X + 1) = Φ1 (X)Φ2 (X). Φ5 (X) = 1 + X + X 2 + X 3 + X 4 .5a).40. . X 3 − 1 = (X − 1)(X 2 + X + 1) = Φ1 (X)Φ3 (X). Cyclotomic polynomials can be computed recursively using Equation (1. Φ7 (X) = 1 + X + X 2 + X 3 + X 4 + X 5 + X 6 . For example.5b) Xn − 1 Φn (X) = ∏ . 1. So the Eisenstein Test can be applied here with s = 1 to show that Φp (X) is irreducible in Z[X]. .41. X 6 − 1 = (X − 1)(X + 1)(X 2 + X + 1)(X 2 − X + 1) = Φ1 (X)Φ2 (X)Φ3 (X)Φ6 (X). .
13
. φ(1) = 1. Of course. .
dn
where the product is taken over all the positive divisors of n. X 12 − 1 = (X − 1)(X + 1)(X 2 + X + 1)(X 2 + 1)(X 2 − X + 1)(X 4 − X 2 + 1) = Φ1 (X)Φ2 (X)Φ3 (X)Φ4 (X)Φ6 (X)Φ12 (X). X 5 − 1 = (X − 1)(X 4 + X 3 + X + 1) = Φ1 (X)Φ5 (X). These are examples of the cyclotomic polynomials Φn (X) ∈ Z[X] which are deﬁned for all n 1 by ∏ (1. As examples we have the irreducible polynomials Φ2 (X) = 1 + X.4) Φp (X) = (X − 1)p−1 + cp−2 (X − 1)p−2 + · · · + c1 (X − 1) + c0
with cr ≡ 0 (mod p) and c0 = p. Φ3 (X) = 1 + X + X 2 . Definition. If we know Φk (X) for k < n. then (1. Φd (X)
dn d<n
The degree of Φn (X) involves a function of n probably familiar from elementary Number Theory.giving (1. X 4 − 1 = (X − 1)(X + 1)(X 2 + 1) = Φ1 (X)Φ2 (X)Φ4 (X). In particular. The Euler function φ : N −→ N is deﬁned by φ(n) = number of k = 1. n for which gcd(n. Φ11 (X) = 1 + X + X 2 + X 3 + X 4 + X 5 + X 6 + X 7 + X 8 + X 9 + X 10 . k) = 1 = (Z/n)×  = number of units in Z/n = number of generators of the cyclic group Z/n. Example. 1. if p 2 is a prime then φ(p) = p − 1.
. the formulae of (1.6) for φ. Remark. ∑ (1. 1. giving o ∑ n (1.
Thus if n = pr1 · · · prs where p1 < p2 < · · · < ps are the prime factors of n and rj > 0. the cyclotomic polynomial Φn (X) is irreducible in Q[X] and hence in Z[X].5) lead to (1.
So for example.
dn
Notice that we can inductively determine φ(n) using this equation.
k ζn = e2πik/n
(0
k
n − 1. Hence. 1.n)=1
Proof.7) φ(mn) = φ(m)φ(n). (1. ∏ t Φn (X) = (X − ζn ).10) φ(n) = µ(d) . For those who know about the M¨bius function µ (which takes values 0. then using standard properties of µ.e. . if p is a prime and r > 0. n) = 1. The complex roots of Φn (X) are the primitive nth roots of unity. d
dn
Similarly. the latter can be used to solve Equation (1. s 1
Furthermore. q are distinct primes.8) φ(n) = φ(pr1 ) · · · φ(prs ). It is also true that whenever m. ±1) o and M¨bius inversion. (X q − 1)(X p − 1)
Recall that an element ζ of a ﬁeld K is a primitive nth root of unity if min{k : 1 k and ζ k = 1} = n. Theorem.. 1. then s 1 (1. We will give a reformulation and proof of this in Theorem 6. n are coprime.6) φ(d) = n. Φpq (X) = (X pq − 1)µ(1) (X pq/p − 1)µ(p) (X pq/q − 1)µ(q) (X pq/pq − 1)µ(pq) = (X pq − 1)(X q − 1)−1 (X p − 1)−1 (X − 1) = (X pq − 1)(X − 1) . gcd(k. when gcd(m.
and the number of these is deg Φn (X) = φ(n).. .9) φ(pr ) = (p − 1)pr−1 . .43..42. n − 1.
14
. For example. Then every k complex nth root of unity has the form ζn = e2πik/n for k = 0.n−1 gcd(t.
t=1. if p. For each n 1.11) Φn (X) =
∏
dn
(X n/d − 1)µ(d) . then (1.
We think of ζn = e2πi/n as the standard complex primitive nth root of unity.2..
Notice that as a result.It can be shown that for each natural number n.. i. n) = 1). if p and q are distinct primes. then φ(pq) = pq − (φ(p) + φ(q) + φ(1)) = pq − (p − 1) − (q − 1) − 1 = (p − 1)(q − 1). φ(n) is even when n > 2.
46. Let θG (d) denote the number of elements in G of order d. This leads to a useful group theoretic result.12). each of order d. h(X) ∈ R[X] with deg g(X) > 0 and deg h(X) > 0.1.12)
∑
dG
θG (d) =
∑
dG
φ(d).
Recall the Euler φfunction satisﬁes Equation (1.45. Since ∪ G= {g ∈ G : g = d}. Hence θG (d) = φ(d) for all d. we say that f (X) has a proper factorisation over R if f (X) = g(X)h(X) for some g(X).6). the number of such elements of C is φ(n).
dG dG
Now if θG (d) < φ(d) for some d. As there are at most d such elements g in G.44. 1. By Lagrange’s Theorem. The above results for polynomials over Q and Z have analogues over the ﬁeld of fractions k(T ) and polynomial ring k[T ]. Example. θG (d) = 0 unless d divides G.
dG
Combining these we obtain (1. hence there must be an element of order G.
15
.
dG
we have G =
∑
dG
θG (d).
It is also worth recording a related general result on cyclic groups. So ∑ ∑ θG (d) φ(d). for each element g ∈ G of order d. A polynomial f (X) ∈ k[T ][X] is an element of k(T )[X]. Lemma. hence ∑ G = φ(d).
Let d be a divisor of G. 2 2
5 Φ6 (X) = X 2 − X + 1 = (X − ζ6 )(X − ζ6 ). If R = k[T ] or k(T ). Proposition. In particular. this gives θG (d) φ(d). Let G be a ﬁnite group satisfying the following condition: • For each n 1. Proof. Then an element g r ∈ C is a generator if and only if gcd(r. there are at most n solutions of xn = ι in G.45. Let n 1 and C = ⟨g⟩ be a cyclic group of order n and a generator g. so G is cyclic. Then G is cyclic and in particular is abelian. the cyclic subgroup ⟨g⟩ G has φ(d) generators. By Proposition 1. 1. For n = 6 we have ζ6 = e Then φ(6) = 2 and
2πi/6
=e
πi/3
√ 3 1 = + i. n) = 1. we would have a strict inequality in place of Equation (1. there are φ(G) elements of order G. where k is a ﬁeld.
Cubic polynomials: Cardan’s method. Proposition (Gauss’s Lemma). Let f (X) ∈ k[T ][X] and s ∈ k[T ]. again we state a version which is slightly more general than the usual one which corresponds to the case where s = 0. so we may as well assume that we want to ﬁnd the complex roots of f (X) = X 3 + pX + q ∈ C[X]. Then f (X) has a proper factorisation over k[T ] if and only it has a proper factorisation over k(T ). Proposition (Eisenstein Test).13) If we introduce u ∈ C for which x=u− then x3 + px + q = 0. Then f (X) is irreducible in k(T )[X] and hence also in k[T ][X]. Suppose that x ∈ C is a root of f (X). 3u
( ( p) p )3 +p u− +q =0 u− 3u 3u
16
. methods of ﬁnding roots are very familiar. .47. . where d = deg f (X). Example. • a0 ̸≡ 0 (mod p(T )2 ). Let us consider the cases of cubic (degree 3) and quartic (degree 4) polynomials. i. (1. Choose ai ∈ k[T ] so that f (X) = a0 + a1 (X − s) + · · · + ad−1 (X − s)d−1 + ad (X − s)d . The following 16th century method of ﬁnding roots of cubics is due to Jerˆme Cardan who seems to have obtained some preliminary versions o from Niccol` Tartaglia by somewhat disreputable means! For historical details see [2. . Here is another version of the Eisenstein Test. Let k be a ﬁeld.1. o A monic cubic f (X) = X 3 + a2 X 2 + a1 X + a0 ∈ C[X] can be transformed into one with no quadratic term by a change of variables X −→ X − a2 /3 giving ) ( ) ( a1 a2 2a3 1 2 3 2 g(X) = f (X − a2 /3) = X − a1 − a2 X + a0 − + ∈ C[X]. For monic linear (degree 1) or quadratic (degree 2) polynomials.4.. In practice we
will usually have k = R or k = C. Then the polynomial X n − T is irreducible in k(T )[X].e.49. p . 1. 3]. d − 1. 3 3 27 Clearly ﬁnding the roots of f (X) is equivalent to ﬁnding those of g(X). Let f (X) ∈ k[T ][X]. 1. • ad ̸≡ 0 (mod p(T )).48. Finding roots of complex polynomials of small degree
♠ ♡ ♢ ♣
In this section we work within the complex numbers and take k ⊆ C. Suppose that p(T ) ∈ k[T ] is an irreducible for which the following three conditions hold: • ak ≡ 0 (mod p(T )) for k = 0. 1. .
Solution. 5 − 26 = 5 + 26 √ Now 5 + 26 has the complex cube roots √ √ √ √ √ √ 3 3 3 5 + 26. 5 + 26 ω 2 . Notice that 5 + 26 > 0 and 5 − 26 < 0. 27 Now if we take u to be a cube root of one of the complex numbers √ q 1 4p3 − ± q2 + 2 2 27 we obtain the desired root of f (X) as x = u − p/3u. √ 27 3 q 1 4p 3 3 − + q2 + 2 2 27
1.and so u3 − hence
p3 + q = 0. 27 q2 + 4p3 .15) x=
3
q 1 − + 2 2
√ q2 +
4p3 p − √ . diﬀering by factors of the form ω r for r = 0.14) q + − − q2 + 2 2 27 2 2 27 or more precisely as √ (1. 1. 27u3 p3 = 0.
17
. Notice that we have a choice of 2 values for u3 and for each of these a choice of 3 values for u. Applying the method above. Find the complex roots of the polynomial f (X) = X 3 + 3X − 10 ∈ R[X]. 27
u6 + qu3 − Solving for u3 we obtain √ where equation q2 + q 1 u =− ± 2 2
3
√
4p3 denotes one of the complex square roots of the discriminant of the quadratic 27 U 2 + qU −
p3 = 0. 2 where ω = e2πi/3 is a primitive cube root of 1. Example. 5 + 26 ω. 2 + 4p3 /27) 3 q − (q 4p 4p3 −q + q 2 + 27 it is easy to verify that there are in fact only 3 choices of the root x which we can write symbolically as √ √ √ √ 3 3 3 q 1 4p q 1 4p3 2+ x= − + (1. we also have √ −1 √ . we reduce to the quadratic equation U 2 − 10U − 1 = 0 √ √ √ whose roots are 5 ± 26 ∈ R. However. since ( ) √ √ 4p3 3 4p −q − q 2 + −q − q 2 + 27 1 27 = −27 √ = 2 .50.
8 8 3 2 16 256 3 4 Clearly ﬁnding the roots of f (X) is equivalent to ﬁnding those of g(X).
This can be done by requiring the vanishing of the discriminant (1. 1. namely √
3
5+
√ 1 26 − √ √ = 3 5 + 26
(√
3
√ )2 5 + 26 − 1 √ . The following method of ﬁnding roots of quartics was publicised by Cardan who attributed it to his student Lodovicio Ferrari. A general monic quartic polynomial f (X) = X 4 + a3 X 3 + a2 X 2 + a1 X + a0 ∈ C[X] can be transformed into one with no cubic term by a change of variables X −→ X − a2 /3 giving g(X) = f (X − a3 /4) = ( ) ( ) ( ) 3 2 1 3 1 1 3 4 1 4 2 2 Y + a2 − a3 Y + a − a2 a3 + a1 Y − a2 a3 − a + a1 a3 + a0 . Suppose that x is a root and introduce numbers y.
Notice that if y = p/2 then we would require q = 0 and then f (X) = X 4 + pX 2 + r = (X 2 )2 + p(X 2 ) + r = 0 can be solved by solving Z 2 + pZ + r = 0. (1. Then for Equation (1.16) (2y − p)x2 − qx + (y 2 − r) = (Ax + B)2 .Here we have x = u − 1/u.17) q 2 − 4(2y − p)(y 2 − r) = 0. Then z 2 = x4 + 2x2 y + y 2 = −px2 − qx − r + 2x2 y + y 2 = (2y − p)x2 − qx + y 2 − r. whence x2 = −t ± (Ax + B).16) we have (x2 + t)2 = (Ax + B)2 . we can use the method of solution of cubics to ﬁnd a root y = t say. so we may as well assume that we want to ﬁnd the complex roots of f (X) = X 4 + pX 2 + qX + r ∈ C[X]. Now choose y to make the last quadratic expression in x a square. √ 3 5 + 26
Quartic polynomials: Ferrari’s method.
18
.17) is a cubic in y. z such that z = x2 + y (we will ﬁx the values of these later). √ 3 5 + 26 Notice that one of these is real. Since Equation (1. 2). so the 3 complex roots of f (X) are ) (√ √ 1 3 5 + 26 − √ ω r (r = 0.
Proof. Remark. (ii) For α ∈ Aut(S). Aut(R) and AutR0 (R) form groups under composition of functions. For a ring R with a subring R0 ⊆ R. then α restricts to the identity on R. Aut(R) = {id}. In the case of cubic and quartic polynomials over C we can obtain all the roots by repeatedly taking square or cube roots (or radicals). Consequently such polynomials are said to be solvable by radicals. this is one of the great early successes of this theory. k α(k1) = −(α(1) + · · · + α(1)) if k < 0. The composition α ◦ β of two automorphisms α.e... Definition. i. (ii) If S is a ring containing a core ring R and α ∈ Aut(S). Aut(S) = AutR (S). Proposition. Later we will see that this is not true in general for polynomials of degree at least 5. For an automorphism α of R. (i) For such a core ring R.54. Thus α = id. k = −(1 + · · · + 1) if k < 0. 1. Hence. Proposition. Let R be a ring and R0 ⊆ R a subring. • An automorphism of R over R0 is a ring isomorphism α : R −→ R for which α(r) = r whenever r ∈ R0 . Let R be one of the core rings Z or Z/n with n > 1. 1. α(r) = r for all r ∈ R. α(1) = 1 and a similar argument to that for (i) shows that α(r) = r for all r ∈ R. 1. Then (i) The only automorphism of R is the identity.
19
. The argument for AutR0 (R) is similar. which we write symbolically as √ x = ± −t ± (Ax + B). i. The set of all automorphisms of R over R0 is denoted AutR0 (R). Proof. α(1) + · · · + α(1) if k > 0.53. −k α(0) if k = 0 1 + · · · + 1 if k > 0.52. Hence Aut(R) forms a group under composition.Thus taking the two square roots of the right hand side we obtain 4 values for x.51. • An automorphism of R is a ring isomorphism α : R −→ R. −k 0 if k = 0 =k1. Automorphisms of rings and ﬁelds 1. The set of all such automorphisms is denoted Aut(R). 1.e. The identity function id = idR : R −→ R is an automorphism.5. β : R −→ R is also an automorphism of R as is the inverse of α. every element has the form k1 for some k ∈ Z.
55. Let k be one of the prime ﬁelds Q or Fp with p > 0 prime.56. Then the induced homomorphism gives an automorphism α∗ : Fr(D) −→ Fr(D). Proof. 1. α −→ α∗ α −→ α∗
is a monomorphism but it is not an epimorphism since there is an automorphism γ : Q(X) −→ Q(X). Let k be a ﬁeld.
20
. The ﬁeld of fractions of the ring of integers Z is the ﬁeld of rationals Q.e. 1. the induced homomorphism α∗ : Fr(D) −→ Fr(D) exists and we need to show it has an inverse. 1.56 need not be an epimorphism. Aut(R) = Autk (R). Example. Let D be an integral domain and α : D −→ D be an automorphism. Given α. Recalling Deﬁnition 1. The homomorphism ( )∗ : Aut(Q[X]) −→ Aut(Q(X)). Combining this example with Proposition 1.58. t) : t ∈ k. we can apply Corollary 1. i. then every automorphism of R restricts to the identity on k.20 to show that (α−1 )∗ ◦ (α)∗ = id = (α)∗ ◦ (α−1 )∗ .57. Hence (α)∗ is invertible with inverse (α−1 )∗ . Example.
1. a11 a22 − a12 a21 ̸= 0 a21 a22 The scalar matrices form a normal subgroup Scal2 (k) = {diag(t. If R is a ring containing k as a subring. α −→ α∗ . There is a monomorphism of groups ( )∗ : Aut(D) −→ Aut(Fr(D)). Here we take D = Q[X] and Fr(Q[X]) = Q(X).1. The group of invertible 2 × 2 matrices over k is the 2 × 2 general linear group over k. {[ ] } a11 a12 GL2 (k) = : aij ∈ k.54(ii) we obtain another useful result. The inverse automorphism α−1 : D −→ D also gives rise to an induced homomorphism (α−1 )∗ : Fr(D) −→ Fr(D). The quotient group is called the 2 × 2 projective general linear group over k. PGL2 (k) = GL2 (k)/ Scal2 (k). γ(f (X)) = f (1/X)
which sends X ∈ Q[X] ⊆ Q(X) to 1/X ∈ Q[X] and so does not restrict to an automorphism of / Q[X].. The homomorphism ( )∗ : Aut(Z) −→ Aut(Q).36. Proposition. we have an example which shows that the monomorphism of Corollary 1.59. Corollary. Proposition. t ̸= 0} ▹ GL2 (k). Since α−1 ◦ α = id = α ◦ α−1 . is an isomorphism and hence Aut(Q) = {id}.
this is also an epimorphism and we leave the proof of this as an exercise. We begin by choosing the element (a11 X + a12 )/(a21 X + a22 ) ∈ k(X) and then using Theorem 1. b ∈ k. (i) We begin by showing that to each invertible matrix [ ] a11 a12 A= ∈ GL2 (k) a21 a22 there is an associated automorphism αA : k(X) −→ k(X). Using the same line of argument as in the proof of Proposition 1. 0 1 1. Using the inverse matrix [ ] a−1 −a−1 b −1 A = 0 1 we similarly obtain a homomorphism αA−1 : k[X] −→ k[X] for which αA−1 (X) = a−1 X − a−1 b. By applying ( )∗ to this we obtain a homomorphism (known as a fractional linear transformation) αA : k(X) −→ k(X) for which αA (X) = a11 X + a12 . a21 X + a22
21
. = Proof.61. Autk (k[X]) ∼ Aﬀ 1 (k). = Proof.22(i) to obtain a homomorphism αA : k[X] −→ k[X] with αA (X) = aX + b. Then Autk (k[X]) and hence Autk (k(X)). {[ ] } a b Aﬀ 1 (k) = : a. αA2 A1 = αA1 ◦ αA2 . Example.Notice that GL2 (k) has another interesting subgroup called the aﬃne subgroup.60. Example. For this we use the element aX + b ∈ k[X] together with the extension result of Theorem 1. It is straightforward to check that for A1 .22(i) to obtain a homomorphism k[X] −→ k(X) that sends X to (a11 X+a12 )/(a21 X+a22 ). A2 ∈ Aﬀ 1 (k). (note the order!) hence there is a homomorphism of groups Aﬀ 1 (k) −→ Autk (k[X]). Let k be a ﬁeld and X an indeterminate. (ii) In fact. We begin by showing that to each aﬃne matrix [ ] a b A= ∈ Aﬀ 1 (k) 0 1 there is an associated automorphism αA : k[X] −→ k[X].55 (or doing a direct calculation) we see that αA−1 is the inverse of αA an so αA ∈ Autk (k[X]). Let k be a ﬁeld and X an indeterminate. 1. Then (i) Autk (k(X)) contains a subgroup isomorphic to PGL2 (k). A −→ αA−1 . a ̸= 0 GL2 (k). Composing with ( )∗ we see that there is a monomorphism Aﬀ 1 (k) −→ Autk (k(X)). In fact. contains a subgroup isomorphic to Aﬀ 1 (k).
which is easily seen to be a monomorphism. In fact. Autk (k(X)) ∼ PGL2 (k).
where α(X) = f (X).t) (X) = tX = X. if deg p(X) > deg q(X) then we can write p0 (X) . But this can only happen if the function f is injective which means that all of these numbers must be 1. but an easy argument then shows the general case. hence ord f = 1. 1.t) is the identity function. The only automorphism of the ﬁeld R is the identity function. There is an associated rational (hence meromorphic) function f given by z −→ f (z). In fact it is easy to see that Scal2 (k)▹GL2 (k) is the kernel of this homomorphism. Also. Therefore passing to the quotient PGL2 (k) = GL2 (k)/ Scal2 (k) we obtain a monomorphism PGL2 (k) −→ Autk (k(X)). (ii) To show that every automorphism of k(X) is a fractional linear transformation is less elementary. Proposition. f (X) = p1 (X) + q(X) where p0 (X). Thus aX + b f (X) = ̸= constant cX + d [ ] a b and the matrix must be invertible.
22
. deﬁned on C with the poles of f deleted. q Now it is easy to see that since α is invertible so is the function f . actually this argument can be modiﬁed to work for any algebraically closed ﬁeld.Again we ﬁnd that αA2 A1 = αA1 ◦ αA2 . There is an associated homomorphism of groups GL2 (k) −→ Autk (k(X)) sending A to αA . Now let c ∈ C. Let α ∈ AutC (C(X)). q(X) ∈ C[X] have no common factors of positive degree. t
−1
showing that αdiag(t. We give a sketch proof for the case of k = C. There is one case where Scal2 (k) is the trivial group. Finally. t). αdiag(t. namely k = F2 . if deg p(X) deg q(X) then the number of poles of f counted with algebraic multiplicity is also ord f . However. deg q(X)}. the real numbers.62. then the order of f (X) is ord f = max{deg p(X). hence Aut(R) = {id}. Now we turn to a more familiar ﬁeld R. c d Clearly not every fractional linear transformation αA : k(X) −→ k(X) maps polynomials to polynomials so ( )∗ : Autk (k[X]) −→ Autk (k(X)) is not an epimorphism. If we write f (X) = p(X) q(X)
where p(X). this is not an injection in general since for each scalar matrix diag(t. Then the number of poles of f counted with algebraic multiplicity is p0 deg p1 (X) + ord . Then the number of solutions counted with algebraic multiplicity of the equation f (z) = c turns out to be ord f . p1 (X) ∈ C[X] and deg p0 (X) < deg q(X).
−δ < y − x < δ) we have −δ = α(−δ) < α(y) − α(x) < α(δ) = δ.
Exercises on Chapter 1 1. Let R be a ring.63.57.e. We recall from Analysis that the rational numbers are dense in the real numbers in the sense that each r ∈ R can be expressed as a limit r = limn→∞ qn . Hence. 1. Thus every automorphism of R is continuous function which ﬁxes all the rational numbers. ( ) ∈ AutR (C). ( )}. What is the relationship between char S and char R ?
23
. in fact..1. hence α(y) − α(x) < δ ε. ε. so there are many elements of order 2 in the group Aut(C). However. It is easy to see that complex conjugation ( ) : C −→ C is an automorphism of C and ﬁxes every real number.. So α preserves order and ﬁxes rational numbers. Note that given an automorphism α ∈ Aut(C). i.
n→∞ n→∞
We will show that an automorphism α ∈ Aut(R) is continuous.
Now for α ∈ Aut(R) and s ∈ R. AutR (C) = {id. 1. First we note that Q ⊆ R is a subring and if α ∈ Aut(R) then α(q) = q for q ∈ Q by Example 1. x < y =⇒ α(y) − α(x) = α(t)2 for some nonzero t ∈ R =⇒ α(x) < α(y). Show that {n ∈ Z : n > 0 and n1 = 0} = {n ∈ Z : n > 0 and nr = 0 for all r ∈ R}. This shows that α is continuous at x. Remark. Let R be an integral domain. First recall that for x. it is not true that every α ∈ Aut(C) ﬁxes every real number! The automorphism group Aut(C) is actually enormous but it is hard to ﬁnd an explicit element other than id and ( ). its value at r depends on its values on Q since f (r) = f ( lim qn ) = lim f (qn ).Proof. (a) Show that every subring S ⊆ R is also an integral domain. the composition α ◦ ( ) ◦ α−1 is also self inverse. Then for a continuous function f : R −→ R.e. If we try to determine Aut(C) the answer turns out to be much more complicated. Deduce that if char R > 0 then these sets are nonempty and char R = min{n ∈ Z : n > 0 and nr = 0 for all r ∈ R}. Now let x ∈ R and ε > 0. where qn ∈ Q. y ∈ R. we have α(s2 ) = α(s)2 . hence it must be the identity function. x<y ⇐⇒ 0<y−x ⇐⇒ y − x = t2 for some nonzero t ∈ R.2. Then we can choose a rational number q such that 0 < q Taking δ = q we ﬁnd that for y ∈ R with y − x < δ (i.
(b) If R is a ﬁeld, give an example to show that a subring of R need not be a ﬁeld. 1.3. For each of the following rings R, ﬁnd the characteristic char R and the characteristic subring of R. Determine which of these rings is an integral domain. In (b) and (c), A is an arbitrary commutative ring. (a) Any subring R ⊆ C. (b) The polynomial ring R = A[X]. (c) The ring of n × n matrices over A, a11 . . . a1n . .. . . R = Matn (A) = . : aij ∈ A . . . . a . . . ann n1 1.4. If R is a commutative ring with unit containing the prime ﬁeld Fp for some prime p > 0, show that the function φ : R −→ R given by φ(t) = tp , deﬁnes a ring homomorphism. Give examples to show that φ need not be surjective or injective. 1.5. Let R and S be rings with unity and Q ▹ S a prime ideal. (a) If φ : R −→ S is a ring homomorphism, show that φ−1 Q = {r ∈ R : φ(r) ∈ Q} ⊆ R is a prime ideal of R. (b) If R ⊆ S is a subring, show that Q ∩ R is a prime ideal of R. (c) If the word ‘prime’ is replaced by ‘maximal’ throughout, are the results in parts (a) and (b) still true? [Hint: look for a counterexample.] (d) If R ⊆ S is a subring and P ▹ R is a maximal ideal, suppose that Q ▹ S is a prime ideal for which P ⊆ Q. Show that Q ∩ R = P . 1.6. Let k be a ﬁeld, R be a ring with unit and let φ : k −→ R be a ring homomorphism. Show that φ is a monomorphism. 1.7. Consider the sets Z(i) = {u + vi : u, v ∈ Z} ⊆ C, Q(i) = {u + vi : u, v ∈ Q} ⊆ C.
(a) Show that Z(i) and Q(i) are subrings of C. Also show that Z(i) is an integral domain, Q(i) is a ﬁeld and Z(i) is a subring of Q(i). (b) Show that the inclusion homomorphism inc : Z(i) −→ Q(i) extends to a monomorphism inc∗ : Fr(Z(i)) −→ Q(i). (c) Show that inc∗ is an isomorphism, so Fr(Z(i)) = Q(i). 1.8. Let R be a commutative ring. (a) If a, b ∈ R, show that there is a unique ring homomorphism ψa,b : R[X] −→ R[X] for which ψa,b (r) = r if r ∈ R and ψa,b (X) = aX + b. If c, d ∈ R, determine ψa,b ◦ ψc,d . If a is a unit, show that ψa,b is an isomorphism and ﬁnd its inverse. (b) Now suppose that R = k is a ﬁeld and a, b ∈ k with a ̸= 0. Prove the following. (i) If f (X) ∈ k[X], the deg ψa,b (f (X)) = deg f (X). (ii) If p(X) ∈ k[X] is a prime then so is ψa,b (p(X)).
24
(iii) If p(X) ∈ k[X] is an irreducible then so is ψa,b (p(X)). 1.9. Let k be a ﬁeld and k[[X]] be the set consisting of all power series
∞ ∑ k=0
ak X k = a0 + a1 X + · · · + ak X k + · · · ,
with ak ∈ k. (a) Show that this can be made into an integral domain containing k[X] as a subring by deﬁning addition and multiplication in the obvious way. ∑ (b) Show that ∞ ak X k ∈ k[[X]] is a unit if and only if a0 ̸= 0. k=0 (c) Show that Fr(k[[X]]) consists of all ﬁnitetailed Laurent series
∞ ∑ k=ℓ
ak X k = aℓ X ℓ + aℓ+1 X ℓ+1 + · · · + ak X k + · · ·
for some ℓ ∈ Z and ak ∈ k. 1.10. Taking k = Q, ﬁnd the quotient and remainder when performing long division of f (X) = 6X 4 − 6X 3 + 3X 2 − 3X − 2 by d(X) = 2X 3 + X + 3. 1.11. Taking k = F3 , ﬁnd the quotient and remainder when performing long division of f (X) = 2X 3 + 2X 2 + X + 1 by d(X) = 2X 3 + 2X. 1.12. Let p > 0 be a prime. Suppose that f (X) = a0 + a1 X + · · · + an X n ∈ Z[X] with p an and that f (X) ∈ Fp [X] denotes the polynomial obtained by reducing the coeﬃcients of f (X) modulo p. If f (X) is irreducible, show that f (X) is irreducible. Which of the following polynomials in Z[X] is irreducible? X 3 − X + 1, X 3 + 2X + 1, X 3 + X − 1, X 5 − X + 1, X 5 + X − 1, 5X 3 − 10X + X 2 − 2. 1.13. Find generators for each of the following ideals: I1 = {f (X) ∈ Q[X] : f (i) = 0} ▹ Q[X], √ I3 = {f (X) ∈ Q[X] : f ( 2) = 0} ▹ Q[X], √ I5 = {f (X) ∈ R[X] : f ( 2 i) = 0} ▹ R[X], √ I2 = {f (X) ∈ Q[X] : f ( 2 i) = 0} ▹ Q[X], √ I4 = {f (X) ∈ R[X] : f ( 2) = 0} ▹ R[X], I6 = {f (X) ∈ R[X] : f (ζ3 ) = 0} ▹ R[X].
1.14. Consider the inclusion inc : Q −→ C and its extension to ε√2 : Q[X] −→ C. Determine the image ε√2 Q[X] ⊆ C. What is ε−√2 Q[X] ⊆ C? Find ker ε√2 ▹Q[X] and ker ε−√2 ▹Q[X]; are these maximal ideals? √ 1.15. Let ω = (−1 + 3i)/2 ∈ C. Consider the inclusion inc : Q −→ C and its extension to εω : Q[X] −→ C. Determine the image εω Q[X] ⊆ C. Determine ker εω ▹Q[X] and decide whether it is maximal. Find another evaluation homomorphism with the same kernel and image. 1.16. Consider the inclusion inc : Q −→ C and its extension to εα : Q[X] −→ C where α is one of the 4 complex roots of the polynomial f (X) = X 4 − 2 ∈ Q[X]. Determine the image εα Q[X] ⊆ C and the ideal ker εα ▹Q[X]; is the latter ideal maximal? What happens if α is replaced by one of the other roots of f (X)? Repeat this problem starting with the inclusion of the real numbers into the complex numbers inc : R −→ C and εα : R[X] −→ C.
25
1.17. Use Cardan’s method to ﬁnd the complex roots of the polynomial f (X) = X 3 − 9X 2 + 21X − 5. 1.18. Consider the real numbers √ √ √ √ 3 3 α = 10 + 108 + 10 − 108, √ β=
3
2 1+ 3
√
√ 7 + 3
3
2 1− 3
√
7 . 3
Find rational cubic polynomials f (X) and g(X) for which f (α) = 0 = g(β). Hence determine these real numbers. 1.19. Prove the ﬁnal part of Example 1.60 by showing that there is an isomorphism of groups Aﬀ 1 (k) ∼ Autk (k[X]). = 1.20. Let k be any ﬁeld. Consider the 6 automorphisms αj : k(X) −→ k(X) (j = 1, . . . , 6) deﬁned by α1 (f (X)) = f (X), α4 (f (X)) = f ((X − 1)/X), α2 (f (X)) = f (1 − X), α5 (f (X)) = f (1/(1 − X)), α3 (f (X)) = f (1/X), α6 (f (X)) = f (X/(X − 1)). Autk (k(X)) isomorphic to
Show that the set consisting of these elements is a subgroup Γk the symmetric group S3 . When k = F2 , show that Γk ∼ GL2 (k). = 1.21. Determine the cyclotomic polynomial Φ20 (X). 1.22. Let p > 0 be a prime. (a) Show that for k
1, the cyclotomic polynomial Φpk (X) satisﬁes Φpk (X) = Φp (X p
k−1
)
and has as its complex roots the primitive pk th roots of 1. (b) Show that Φpk (X) ∈ Q[X] is irreducible. (c) Generalize part (a) to show that if n = pr1 · · · prk is the prime power factorization of n 1 k with the pi being distinct primes and ri > 0, then Φn (X) = Φp1 ···pk (X p1 1.23. For n 2, show that X φ(n) Φn (X −1 ) = Φn (X).
−1 1.24. Show that for n 1, ζn + ζn = 2 cos(2π/n). −1 −2 2 Find expressions for ζ5 +ζ5 and ζ5 +ζ5 in terms of cos(2π/5). Hence ﬁnd a rational polynomial which has cos(2π/5) as a root.
r −1 r1 −1 ···pkk
).
1.25. Let p > 0 be a prime and K be a ﬁeld with char K = p. (a) Show that if ζ ∈ K is a pth root of 1 then ζ = 1. Deduce that if m, n > 0 and p n, then every npm th root of 1 in K is an nth root of 1. (b) If a ∈ K, show that the polynomial X p − a ∈ K[X] has either no roots or exactly one root in K.
26
v ∈ Q not both zero. L is also said to be an extension (ﬁeld ) of K. however R and C are uncountable. If the elements 1. We have C = {x + yi : x. if K ̸= L. 2 clearly span the Qvector space Q( 2). M
2. √ Show that [Q( 2) : Q] = 2. √ 2. If we have two extensions L/K and M/L then it is a straightforward to verify that K and so we have another extension M/K. Now recall that √ √ √ 2 ∈ Q. in fact it is easy to see that we would then also have u. An extension of ﬁelds L/K is ﬁnite (dimensional ) if [L : K] < ∞. Hence 1. so 1.4. Consider the extension Q( 2)/Q where √ √ Q( 2) = {x + y 2 : x. Definition. We write K L or L/K to indicate this. these elements are also linearly independent / over R and therefore they form a basis. We will call dimK L the degree or index of the extension L/K and use the notation [L : K] = dimK L.5. x ∈ L). √ √ Solution. Given two extensions L/K and M/L. Thus we would have √ u 2 = − ∈ Q. The elements 1.1. we say that L/K is a subextension of M/K and sometimes write L/K M/K. y ∈ Q}. v √ √ which we know to be false.1. Definition. Show that the extension C/R is ﬁnite.
2. Fields and subﬁelds 2. v both nonzero.2. whence [C : R] = 2. Definition. Solution. 2 were linearly dependent we would have u + v 2 = 0 for some / u.3.e. Since i ∈ R. 2. Example.CHAPTER 2
Fields and their extensions
2.
27
. while R/Q and C/Q are both inﬁnite. otherwise it is inﬁnite (dimensional ). Let K and L be ﬁelds and suppose that K ⊆ L is a subring. The inﬁniteness of R/Q and C/Q are consequences of the fact that any ﬁnite dimensional vector space over Q is countable.. and write K < L if K is a proper subﬁeld of L. i. An important fact about an extension of ﬁelds L/K is that L is a Kvector space whose addition is the addition in the ﬁeld L while scalar multiplication is deﬁned by u · x = ux (u ∈ K. Example. Then we say that K is a subﬁeld of L. i span C as a vector space over R. 2 are linearly independent and so form a basis for Q( 2) √ over Q and [Q( 2) : Q] = 2. y ∈ R}. A basis for the Qvector space R is known as a Hamel basis.
2.6. Theorem. Let L/K be a subextension of M/K. (i) If one or both of the dimensions [L : K] or [M : L] is inﬁnite then so is [M : K]. (ii) If the dimensions [L : K] and [M : L] are both ﬁnite then so is [M : K] and [M : K] = [M : L] [L : K]. Proof. (i) If [M : K] is ﬁnite, choose a basis m1 , . . . , mr of M over K. Now any element u ∈ M can be expressed as u = t1 m1 + · · · + tr mr , where t1 , . . . , tr ∈ K; but since K ⊆ L, this means that m1 , . . . , mr spans M over L and so [M : L] < ∞. Also L is a Kvector subspace of the ﬁnite dimensional Kvector space M , hence [L : K] < ∞. (ii) Setting r = [L : K] and s = [M : L], choose a basis ℓ1 , . . . , ℓr of L over K and a basis m1 , . . . , ms of M over L. Now let v ∈ M . Then there are elements y1 , . . . , ys ∈ L for which v = y1 m1 + · · · + ys ms . But each yj can be expressed in the form yj = x1j ℓ1 + · · · + xrj ℓr for suitable xij ∈ K. Hence, v= ( r s ∑ ∑
j=1 i=1
) xij ℓi mj =
s r ∑∑ j=1 i=1
xij (ℓi mj ),
where each coeﬃcient xij is in K. Thus the elements ℓi mj (i = 1, . . . , r, j = 1, . . . , s) span the Kvector space M . Now suppose that for some tij ∈ K we have
r s ∑∑ j=1 i=1
tij (ℓi mj ) = 0.
On collecting terms we obtain
( r s ∑ ∑
j=1 i=1
) tij ℓi mj = 0,
where each coeﬃcient means that for each j,
∑r
i=1 tij ℓi
is in L. By the linear independence of the mj over L, this
r ∑ i=1
tij ℓi = 0.
By the linear independence of the ℓi over K, each tij = 0. Hence the ℓi mj form a basis of M over K and so [M : K] = rs = [M : L] [L : K].
28
We will often indicate subextensions in diagrammatic form where larger ﬁelds always go above smaller ones and the information on the lines indicates dimensions MF
[M :L]
A
L 1
[L:K]
6 &
[M :K]=[M :L] [L:K]
K
!
We often suppress ‘composite’ lines such as the dashed one. Such towers of extensions are our main objects of study. We can build up sequences of extensions and form towers of arbitrary length. Thus, if L1 /K, L2 /L1 , . . . , Lk /Lk−1 is a such a sequence of extensions, there is a diagram Lk
Lk−1
L1
K 2.2. Simple and ﬁnitely generated extensions 2.7. Definition. Let F be a ﬁeld and K F . Given elements u1 , . . . , ur ∈ F we set ∩ K(u1 , . . . , ur ) = L
K L F u1 ,...,ur ∈L
which is the smallest subﬁeld in F that contains K and the elements u1 , . . . , ur . The extension K(u1 , . . . , ur )/K is said to be generated by the elements u1 , . . . , ur ; we also say that K(u1 , . . . , ur )/K is a ﬁnitely generated extension of K. An extension of the form K(u)/K is called a simple extension of K with generator u. We can extend this to the case of an inﬁnite sequence u1 , . . . , ur , . . . in F and denote by K(u1 , . . . , ur , . . .) F the smallest extension ﬁeld of K containing all the elements ur . It can be shown that (2.1) K(u1 , . . . , ur ) = } { f (u1 , . . . , ur ) ∈ F : f (X1 , . . . , Xr ), g(X1 , . . . , Xr ) ∈ K[X1 , . . . , Xr ], g(u1 , . . . , ur ) ̸= 0 . g(u1 , . . . , ur ) Reordering the ui does not change K(u1 , . . . , un ).
29
2.8. Proposition. Let K(u)/K and K(u, v)/K(u) be simple extensions. Then K(u, v) = K(u)(v) = K(v)(u). More generally, K(u1 , . . . , un ) = K(u1 , . . . , un−1 )(un ) and this is independent of the order of the sequence u1 , . . . , un . 2.9. Theorem. For a simple extension K(u)/K, exactly one of the following conditions holds. (i) The evaluation at u homomorphism εu : K[X] −→ K(u) is a monomorphism and on passing to the fraction ﬁeld gives an isomorphism (εu )∗ : K(X) −→ K(u). In this case, K(u)/K is inﬁnite and u is said to be transcendental over K. (ii) The evaluation at u homomorphism εu : K[X] −→ K(u) has a nontrivial kernel ker εu = (p(X)) where p(X) ∈ K[X] is an irreducible monic polynomial of positive degree and the quotient homomorphism εu : K[X]/(p(X)) −→ K(u) is an isomorphism. In this case K(u)/K is ﬁnite with [K(u) : K] = deg p(X) and u is said to be algebraic over K. Proof. (i) If ker εu = (0), all that needs checking is that (εu )∗ is an epimorphism; but as u is in the image of (εu )∗ this is obvious. (ii) When ker εu ̸= (0), Theorem 1.31(iv) implies that the image of εu is a subﬁeld of K(u) and since it contains u it must equal K(u). Hence εu is an isomorphism. Using Long Division, we ﬁnd that every element of K[X]/(p(X)) can be uniquely expressed as a coset of the form f (X) + (p(X)), where deg f (X) < deg p(X). Hence every element of K[X]/(p(X)) can be uniquely expressed as a linear combination over K of the d cosets 1 + (p(X)), X + (p(X)), X 2 + (p(X)), . . . , X d−1 + (p(X)), where d = deg p(X). Via the isomorphism εu under which εu (X k + (p(X))) = uk , we see that the elements 1, u, . . . , ud−1 form a basis for K(u) over K. √ √ √ √ 2.10. Example. For the extension Q( 2, 3)/Q we have [Q( 2, 3) : Q] = 4. √ Proof. By Example 2.4 we know that [Q( 2) : Q] = 2. We have the following tower of extensions. √ √ Q( 2, 3)
√ √ √ [Q( 2, 3):Q( 2)]
√ Q( 2)
2
√ √ √ √ √ [Q( 2, 3):Q]=2[Q( 2, 3):Q( 2)]
Q √ √ √ We will show that [Q( 2, 3) : Q( 2)] = 2.
30
pn ) : Q( p1 . Proposition. 3 are linearly independent over Q( 2) √ √ √ √ √ and therefore form a basis of Q( 2. . . So 1. pn ) : Q] = 2n . / √ √ √ √ √ √ Hence [Q( p1 . Using the formula √ √ √ √ [Q( 2. . b ∈ Q( 2).
31
. √ √ 2. Writing √ √ 3=v+w 2 with v. There are some other subﬁelds of Q( 2. 3 span Q( 2. while v. 3). . . i) : Q] = 4. 2. .13. . 3) : Q] = 4. . √ √ √ / Proof. This shows that [Q( 2. . we ﬁnd that √ v 2 + 2w2 + 2vw 2 = 3 ∈ Q. . 3)
2
√ Q( 2)
r rrr rrr rrr
2
vvv vvv vvv 2 vvv
√ Q( 3)
2
vvv vvv2 vvv v
Q
r rrr rrr rr 2 rrr
√ Q( 6)
One idea in the veriﬁcation of Example 2. √ we obtain [Q( 2. w ∈ Q. w ̸= 0 √ √ √ would implies that 2 ∈ Q which is false. Then √ √ √ pn ∈ Q( p1 . . The possibilities v = 0 or w = 0 are easily ruled out. i) : Q( 2)] [Q( 2) : Q]. √ √ √ √ Since i2 + 1 = 0. . pn−1 )] = 2 and [Q( p1 . √ and hence 2vw 2 ∈ Q. we have Q( 2.10 can be extended to provide a useful general result whose proof is left as an exercise. Remark.√ √ √ √ √ √ Notice that if u ∈ Q( 2. 3) which are conveniently displayed in the following diagram. . √ √ √ √ √ √ so 1. 3) = Q( 2)( 3) then u = a + b 3 for some a. But if these are linearly dependent then 3 ∈ Q( 2). . i) : Q] = [Q( 2. 3) over Q( 2). i ∈ Q( 2) since i is not real and Q( 2) R. pn−1 ). √ √ 2. i) = Q( 2)(i) and [Q( 2. i) : Q] = 4. Example. pn be a sequence of distinct primes pi > 0. For the extension Q( 2. . 3) : Q( 2)] = 2 and so √ √ [Q( 2. Let p1 . . √ √ Q( 2. Also. . .11. . We know that [Q( 2) : Q] = 2. i)/Q we have [Q( 2.12. i) : Q( 2)] = 2. .
i) ∩ R being a subﬁeld of R. whence Emn Q(21/m . By Theorem 2.
32
. n are coprime.
1. Example. 21/n ). let En = Q(21/n )
R.√ √ This example also has several other subﬁelds. Solution. with only Q( 2) = Q( 2. 21/m = (21/n )n/m ∈ En . As gcd(m. (ii) Since n/m is an integer. Emn . Combining these inclusions we obtain Emn = Q(21/m . 21/n ). and the induced homomorphism ε21/n : Q[X]/(X n − 2) −→ En is an isomorphism. hence Q(21/m .38 using the prime 2 to the polynomial X n − 2 ∈ Z[X]. 21/n ). (ii) If m 1 with m  n. 21/n ) there are integers r.6 we have n = [En : Q] = [En : Em ] [Em : Q] = m[En : Em ]. (iii) By (ii) we have Em Emn and En Emn . show that Em En and determine [En : Em ]. we ﬁnd that ker ε21/n = (X n − 2) ▹ Q[X]. i)
2
uu uu 2 uu uu u
uu uu uu u 2 uuu u
Q(i)
2
Q
rr rr rr r rr 2 rr
√ Q( 2 i)
2. C
2
R
∞
∞
√ Q( 2)
tt tt tt t tt
2
√ Q( 2. For n nth root of 2. (iii) If m. show that Emn = Q(21/m . Hence [En : Q] = n. where 21/n ∈ R denotes the positive real
(i) Show that [En : Q] = n. s for which rm + sn = 1 and so 1 rm + sn r s = = + . Applying the Eisenstein Test 1. so Em = Q(21/m ) ⊆ En . n) = 1. 21/n ). mn mn n m This shows that 21/mn = (21/n )r (21/m )s ∈ Q(21/m . whence [En : Em ] = n/m. (i) Consider the evaluation homomorphism ε21/n : Q[X] −→ En .14.
(f) Is i ∈ Q(ζ7 )? 2. (b) Show that cos(2π/5). 4 are roots of the polynomial f (X) = 16X 5 − 20X 3 + 5X − 1 = (X − 1)(4X 2 + 2X − 1)2 and deduce that [Q(cos(2π/5)) : Q] = 2.Exercises on Chapter 2 √ √ 2. (a) Explain why [Q(ζ5 ) : Q] = 4. then show that u. (c) Show that for t ∈ R. Find the three subﬁelds L Q( 3. Show that [Q( 3.1. 1. b ∈ K are distinct. Let ζ5 = e2πi/5 ∈ C. sin(2π/5) i ∈ Q(ζ5 ). 6 are roots of the polynomial f (X) = 64X 7 − 112X 5 + 56X 3 − 7X − 1 = (X − 1)(8X 3 + 4X 2 − 4X − 1)2 and deduce that [Q(cos(2π/7)) : Q] = 3.6.12 by induction on n. 2. sin(2π/7) i ∈ Q(ζ7 ). What is [Q(sin(2π/7) i) : Q]? (e) Display the relationship between the ﬁelds Q. (c) Show cos 7t = 64 cos7 t − 112 cos5 t + 56 cos3 t − 7 cos t.14. i) : Q] = 4. Show that [Q( p.3. v ∈ L satisfy u2 = a and v 2 = b. This question is for those who like lots of calculation or using Maple. 2. . Let K a ﬁeld with char K ̸= 2 and suppose that L/K is an extension. cos 5t = 16 cos5 t − 20 cos3 t + 5 cos t.4. 1. and Q(ζ5 ) in a suitable diagram. q > 0 be distinct primes. (b) Show that cos(2π/7).
33
. . In this question we continue to consider the situation described in Example 2. Show that the extension Q( p)/Q has [Q( p) : Q] = 2. . 2. 2. q) : Q( p)] = 2.5. Q(sin(2π/7) i) and Q(ζ7 ) in a diagram. . suppose that u. Show that [Q(i) : Q] = 2. [Hint: ﬁrst show that u±v ̸= 0 and deduce that u−v ∈ K(u+v). Show that K(u.2. 3. Let ζ7 = e2πi/7 ∈ C.8. v ∈ K(u+v). Let p.9. (d) Show that the numbers cos(2kπ/5) with k = 0. (e) Display the relationship between the ﬁelds Q.] 2. √ √ 2. (a) Explain why [Q(ζ7 ) : Q] = 6.7. Q(cos(2π/5)). Prove Proposition 2. i) with [L : Q] = 2 and display their relationship in a diagram. Q(cos(2π/7)). Let p ∈ N be an prime. If a. Show that the numbers cos(2kπ/7) with k = 0. √ √ √ 2. indicating which ones are subﬁelds of R. 2. (d) Show that sin(2π/7) i is a root of g(X) = 64X 7 + 112X 5 + 56X 3 + 7X = X(64X 6 + 112X 4 + 56X 2 + 7) and that 64X 6 + 112X 4 + 56X 2 + 7 ∈ Q[X] is irreducible. v) = K(u + v).
=
where τn ∪ composition order 2. Show that AutQ (E) = {id}. (d) Which of the subﬁelds in part (c) contain the element 21/2 + 21/3 ?
34
.(a) Show that AutQ (En ) =
{id}
if n is odd. τn } ∼ Z/2 if n is even. has (b) Let E = En R.
n 1
(c) Display the 6 subﬁelds of E12 in a diagram.
{id.
9(ii). Example. If t ∈ L is algebraic over K then [K(t) : K] = deg minpolyK. Consider C/Q.1. Proposition.t (X)) ̸= (0).9 allows us to characterize algebraic elements in other ways.4.2.√2 (X) = X 2 − 2.9(ii). (i) t is algebraic over K. 3. If t ∈ L is algebraic over K then by Proposition 3. Proposition. Notice in particular that for an element t ∈ K. 3.
35
√ 2 ∈ C over Q is
.1. We will often use this without further comment. Definition.3.CHAPTER 3
Algebraic extensions of ﬁelds
3.2. Theorem 2. This follows from Theorem 2. Then minpolyK. p(X) = minpolyK. 3.t (X) for some u ∈ K. The degree of minpolyK.t (X)  p(X) and so p(X) = u minpolyK.5. recall the following notion. Let t ∈ L. From Theorem 2. when p(X) is monic. ker εt = (minpolyK. 3. Proof.t (X) is called the degree of t over K and is denoted degK t. The minimal polynomial of minpolyQ. 3. Algebraic extensions Let L/K be an extension of ﬁelds. Suppose that t ∈ L is algebraic over K and that p(X) ∈ ker εt with deg p(X) = deg minpolyK. Definition. (iii) The extension K(t)/K is ﬁnite dimensional. (ii) The evaluation homomorphism εt : K[X] −→ L has nontrivial kernel.6. An element t ∈ L is algebraic over K if there is a nonzero polynomial p(X) ∈ K[X] for which p(t) = 0.t (X).t (X). Remark. the polynomial p(X) = X − t ∈ K[X] satisﬁes p(t) = 0.t (X) = degK t. In particular. where minpolyK. Then the following conditions are equivalent. so t is algebraic over K. 3.t (X) ∈ K[X] is an irreducible monic polynomial called the minimal polynomial of t over K.
Clearly X 2 − 2 ∈ ker ε√2 since ( 2)2 − 2 = 0. Consider C/Q.i (X) = X 2 + 1. 3)
Referring to Example 2. Find the minimal polynomial of √ √ 3 over Q. Consider C/Q. √ minpolyQ(√2). 3.10 or Proposition 2. Since Q( 2 + 3) Q( 2. The minimal polynomial of i ∈ C over Q is X 2 + 1. Recall from Example 1.10 we see that
Let us ﬁnd a nonzero polynomial in ker ε√2+√3 ▹Q[X]. √ √ degQ ( 2 + 3) = 4. Clearly X 2 + 1 ∈ ker εi since i2 + 1 = 0. Example. Then Φ6 (X) ∈ ker εζ6 so minpolyQ. Since Φ6 (X) is irreducible and monic. 3). ζ6 ∈ C over Q.9.4.
2+
Solution. 3. Consider C/Q. As [Q(i) : Q] = 2. Find the minimal polynomial of the primitive 6th root of unity.8. we must have minpolyQ.
36
. hence minpolyQ. 3.44 that ζ6 is a root of the irreducible cyclotomic polynomial Φ6 (X) = X 2 − X + 1. Since this is monic and of degree 2. 3− 2= =√ ( 3 + 2) 2+ 3 So we have √ √ √ ) √ √ √ 1( √ 2= ( 2 + 3) − ( 3 − 2) ∈ Q( 2 + 3). Example. we have minpolyQ.√2 (X) = [Q( 2) : Q] = 2.√ Proof.ζ6 (X)  Φ6 (X). Proof.√2 (X) = X 2 − 2. √ One polynomial in ker ε√2+√3 ▹Q( 2)[X] is √ √ √ √ √ (X − ( 2 + 3))(X − ( 2 − 3)) = X 2 − 2 2X − 1. By Example 2. 3) we must have √ √ √ √ Q( 2 + 3) = Q( 2. 2 √ √ √ √ ) √ √ 1( √ 3= ( 2 + 3) + ( 3 − 2) ∈ Q( 2 + 3).7. Example.ζ6 (X) = Φ6 (X) and so degQ ζ6 = 2. Notice that √ √ √ √ √ √ √ √ ( 3 − 2)( 3 + 2) 1 √ √ √ ∈ Q( 2 + 3).12 we see that 2 + 3 ∈ Q( 2). √ √ √ Referring to Example 2.
√ √ hence Q( 2.√2+√3 (X) = X 2 − 2 2X − 1. hence / √ √ degQ(√2) ( 2 + 3) = 2. 2 √ √ √ √ √ √ Q( 2 + 3). Solution. √ deg minpolyQ.
Proposition. Definition. Then K(u) = L if and only dimK K(u) = dimK L. Then K(u1 .12. un )/K is a ﬁnite extension.−√2+√3 (X) = X 2 + 2 2X − 1. Let L/K be an extension and suppose that u1 . Let t ∈ L be algebraic over K. Later we will see that when char K = 0 every ﬁnite extension L/K has a primitive element.t (X). . Use induction on n together with Proposition 2. To see that q(X) is irreducible. .−√2+√3 (X) √ √ = (X 2 − 2 2X − 1)(X 2 + 2 2X − 1)
√ √ Then p( 2 + 3) = 0 so p(X) ∈ ker εt . Since deg p(X) = 4 and p(X) is monic. we must have q1 (t) = 0 or q2 (t) = 0. Consider
√ minpolyQ(√2).t (X). .
. Sometimes the minimal polynomial of an element in an extension is introduced in a diﬀerent but equivalent way. Let L/K be a ﬁnite extension. Now as q1 (t)q2 (t) = 0. p(X) = minpolyQ(√2). Then u is a primitive element for L/K if and only if degK u = [L : K]. . suppose that q(X) = q1 (X)q2 (X) with deg qi (X) < deg q(X). we have minpolyQ. Since degK u = dimK K(u) and [L : K] = dimK L the result follows.t (X) and therefore q(X) = minpolyK. Lemma. The next Lemma will often be useful.√2+√3 (X) = X 4 − 10X 2 + 1. . 3. K(u) ⊆ L is a ﬁnite dimensional Kvector subspace. 3. 3. Definition. An element u ∈ L for which L = K(u) is called a primitive element for the extension L/K. Proof.14. q(X) = minpolyK. 3.√2+√3 (X) minpolyQ(√2). The extension L/K is algebraic or L is algebraic over K if every element t ∈ L is algebraic over K. We now come to an important notion for extensions.6(ii).13. It is easy to see that I(t) ▹ K[X] and therefore I(t) = (q(X)) for some monic generator q(X). 3. . . These possibilities give q(X)  q1 (X) or q(X)  q2 (X) and so deg q(X) deg q1 (X) or deg q(X) deg q2 (X). Proof.8 and Theorem 2.
37
= X 4 − 10X 2 + 1.Similarly. Then I(t) = {f (X) ∈ K[X] : f (t) = 0} ⊆ K[X] is an ideal which is principal and has an irreducible monic generator q(X) ∈ K[X]. hence q1 (X) ∈ I(t) or q2 (X) ∈ I(t). The irreducible monic polynomial minpolyK. . contradicting the above assumption that deg qi (X) < deg q(X). Proof. In fact.t (X) is in I(t) so q(X)  minpolyK. un ∈ L are algebraic. Let L/K be a ﬁnite extension and u ∈ L.11. Lemma.10.
20. so there is a polynomial p(X) = p0 + p1 X + · · · + pm X m ∈ L[X] of positive degree with p(u) = 0. The sets Calg and Ralg are both countable. Proposition. When dealing with algebraic extensions of Q we will usually work with subﬁelds of Q = Calg . Elements of C−Calg are called transcendental complex numbers. Then the extension M/K is algebraic. v)/K is a ﬁnite dimensional extension. Hence for suitable coeﬃcients cj ∈ K not all zero and some m 1 we have c0 + c1 t + · · · + cm tm = 0. By Lemma 3. the powers 1. so by Proposition 3. u)/K is ﬁnite. . . 3. But this means that t is algebraic over K. . pm . We end this section with a technical result. Let t ∈ L. pm )/K is ﬁnite and so is K(p0 . v) and if u ̸= 0. For an extension L/K. Clearly K ⊆ Lalg . Then by Lemma 3. Definition. Proposition. v) is algebraic over K.6(ii). let Lalg = {t ∈ L : t is algebraic over K} ⊆ L. the extension K(p0 . Let u ∈ M . Proof. Let K(u)/K be a ﬁnite simple extension. In particular. . examples are e and π. Now for any subextension F/K K(u)/K we can also consider minpolyF. By Theorem 2. . 3. pm ). Lalg is a subﬁeld containing K and Lalg /K is algebraic.3. Let M/L and L/K be algebraic extensions.17. . in the extension R/Q the subﬁeld Ralg = Calg ∩ R C
consists of all the real algebraic numbers. Then there are only ﬁnitely many subextensions F/K K(u)/K. hence every element of K(u. . . . Consider the minimal polynomial minpolyK. u−1 is also in K(u. . v). uv and u−1 are all algebraic over K. u + v and uv are in K(u.15. pm . . Let u. Then u is algebraic over L. Example.13.u (X) = c0 + c1 X + · · · + ck−1 X k−1 + X k ∈ F [X]. whereas C and R are uncountable. 3. . t.u (X) ∈ K[X]. . K(u. Proof. tn . . when viewed as elements of this vector space. Proposition. so there are in fact many more transcendental numbers but it can be hard to determine whether a given number is transcendental or not. v ∈ Lalg . . 3. . Proposition. Similarly. Since the Kvector space L is ﬁnite dimensional. A more usual notation for Calg is Q since this is the algebraic closure of Q which will be discussed later. . .13.16. We must show that Lalg L. . Therefore u + v. 3. Then L/K is algebraic. In the extension C/Q we can consider Calg C which is called the subﬁeld of algebraic numbers. Proof. u)/K(p0 . .19. must be linearly dependent over K. K(p0 .18. .
38
. Proof. Let L/K be a ﬁnite extension. u is algebraic over K.15. . For an extension L/K. .
hence minpolyF.u (X) ∈ F0 [X] is irreducible since it is irreducible in F [X].21. Then there is a ﬁnite extension L/K for which p(X) has a root in L. . Let K be a ﬁeld and p(X) ∈ K[X] be a polynomial of positive degree. We have [K(u) : F ] = deg minpolyF. the extension ﬁeld of K generated by the coeﬃcients of minpolyF.22. Is there an extension ﬁeld E/K for which p(X) factorizes into linear factors in E[X]? 3. . .22.25. . . 3. . 3.u (X). We already know how to answer Question 3. .u (X) = [K(u) : F0 ]. Is there an extension ﬁeld L/K for which p(X) has a root in L? A stronger version of this question is the following. if we have such a ﬁeld E then the distinct roots u1 . so there are only a ﬁnite number of possibilities for minpolyF. ck−1 ). p(X) ∈ K[X] splits in E/K or over E if it factorizes into linear factors in E[X]. Question. Definition.21. 3.33 implies that minpolyK. This shows that there are only ﬁnitely many subextensions F/K has the form K(a0 . Now for any j we can form the quotient ﬁeld K[x]/(qj (X)) which is a ﬁnite dimensional (simple) extension of K and in which the coset X + (qj (X)) satisﬁes the equation qj (X + (qj (X))) = 0 + (qj (X)). Then F0 F and so minpolyF. . .which divides minpolyK. Let K be a ﬁeld and p(X) ∈ K[X] be a polynomial of positive degree.
39
K(u)/K. Definition.u (X) = minpolyF0 . a1 .u (X) in K(u)[X]. Proof. . 3. Now consider F0 = K(c0 . . . Of course.u (X) = deg minpolyF0 . We begin by factorizing p(X) ∈ K[X] into irreducible monic factors qj (X) together with a constant factor c: p(X) = cq1 (X) · · · qr (X). . c1 . .24. Such a minimal extension of K is called a splitting ﬁeld of p(X) over K and we will sometimes denote it by K(p(X)) or Kp . where a0 + a1 X + · · · + aℓ−1 X ℓ−1 + X ℓ ∈ K(u)[X] is a factor of minpolyK.u (X). . aℓ−1 ). Splitting ﬁelds and Kronecker’s Theorem We can now answer a basic question. Of course.u (X) has only ﬁnitely many monic divisors in K(u)[X]. Question.23. this construction is only interesting if qj (X) to has degree bigger than 1 since a linear polynomial already has a root in K.2. The Unique Factorization Property 1. uk ) E which is the smallest subﬁeld of E that answers Question 3. 3. uk of p(X) in E generate a subﬁeld K(u1 .u (X). Theorem (Kronecker’s Theorem: ﬁrst version). each of which
. hence F = F0 . Hence p(X) has a root in K[x]/(qj (X)).u (X) in F [X]. .
having found one root u1 in an extension L1 /K we discard the linear factor X − u1 and consider the polynomial p(X) ∈ L1 [X]. We then refer to a subﬁeld of C which is a splitting ﬁeld as the splitting subﬁeld. 3. C √ Q( 2. Theorem (Kronecker’s Theorem: second version).26. 3. i)/Q( 2) which has degree 2. Notice that p(X) = (X 2 − 2)(X 2 + 2). Find a splitting ﬁeld E/Q for p(X) = X 4 −4 over Q and determine [E : Q]. Namely. i) and the extension Q( 2.27. Example. √ √ √ √ so ﬁrst we adjoin the roots ± 2 of (X 2 − 2) to form Q( 2. i) : Q] = 4. X − u1
p1 (X) =
We can repeat the argument to form a ﬁnite extension of L1 (and hence of K) containing a root of p1 (X) and so on. i)
adjoin roots of X 2 + 2 2 ∞
√ Q( 2)
adjoin roots of X 2 − 2 2
Q √ √ Thus the splitting subﬁeld of p(X) over Q in C is Q( 2. The complex roots of X 2 + 2 are ± 2i √ and these are not real. In practise we often have extension ﬁelds ‘lying around in nature’ containing roots and we can work inside of these. i) and [Q( 2. so this polynomial is irreducible in Q( 2)[X]. √ √ Next consider the polynomial X 2 + 2 ∈ Q( 2)[X]. 2i) = Q( 2. Hence we need to consider √ √ √ √ √ Q( 2. Solution. Let K be a ﬁeld and p(X) ∈ K[X] be a polynomial of positive degree. At each stage we either already have another root in L1 or we need to enlarge the ﬁeld to obtain one.
40
.To answer Question 3.22 we iterate this construction. Then there is a ﬁnite extension E/K which is a splitting ﬁeld of p(X) over K. When working over Q (or any other subﬁeld of C) we can always ﬁnd roots in C by the Fundamental Theorem of Algebra. − 2) = Q( 2) which gives an √ extension Q( 2)/Q of degree 2.
. . Let F/K be an extension ﬁeld and p(X) ∈ K[X]. Example. 3. • two distinct roots −b + δ −b − δ = (2a)−1 (−b + δ). i)
adjoin roots of X 2 − 2 2 ∞
√ Q( 2i)
adjoin roots of X 2 + 2 2
Q An important point is that if a splitting ﬁeld exists inside of a given extension ﬁeld F/K. thus giving the tower C √ Q( 2. . .4. 3. But K(u1 . . • one root −b/(2a) = −(2a)−1 b if ∆ = 0. uk ∈ F be the distinct roots of p(X) in F . K(u1 .29. . = (2a)−1 (−b − δ). . 2a 2a if ∆ = δ 2 for some nonzero δ ∈ K. Proposition. Let K be a ﬁeld of characteristic diﬀerent from 2. Proposition. F are
Proof.28. . If E1 . . Find a splitting ﬁeld E/Q for p(X) = X 3 −2 over Q and determine [E : Q]. Since we will frequently encounter quadratic polynomials we record a useful result on roots of such polynomials. uk ) must be contained in any splitting subﬁeld. Then the quadratic polynomial p(X) = aX 2 + bX + c ∈ K[X] has • no roots in K if ∆ is not a square in K. In particular. E2 splitting subﬁelds for p(X) over K then E1 = E2 .. Now √ √ √ 3 3 3 p(X) = (X − 2)(X 2 + 2X + ( 2)2 )
41
. so E1 = K(u1 . . .e. it is unique as a subﬁeld of F . . uk ) is the smallest subﬁeld containing K and all the uj .Of course we could have started by ﬁrst adjoining roots of X 2 + 2 and then adjoining roots of X 2 − 2. . .38. uk ) = E2 .30. Let u1 . . 3. where δ is one of the two square roots of ∆ in some extension of K such as the algebraic closure K which we will introduce in Section 3. the splitting ﬁeld of p(X) over K is K if ∆ is a square in K and K(δ) otherwise. . By deﬁnition. One root of p(X) is √ √ 3 2 ∈ R so we adjoin this to Q to form an extension Q( 3 2)/Q of degree 3. p(X) is irreducible over Q. i. By the Eisenstein Test 1. Recall that p(X) = aX 2 + bX + c ∈ K[X] is quadratic if a ̸= 0 and its discriminant is ∆ = b2 − 4ac ∈ K. The proof of the next result is the standard one which works provided 2 has an inverse in K. Solution. when char K ̸= 2.
31. If F/K is ﬁnite. Definition. F ) is closed under composition but is not always a group since elements are not necessarily invertible. F ) = {u ∈ F : p(u) = 0}. 3. √ √ 2 An alternative strategy would have been to adjoin one of the other roots 3 2 ζ3 or 3 2 ζ3 ﬁrst. We will also use the following notation. For extensions F/K and L/K. ζ3 )/Q
3.
42
.3. which happens precisely when p(X) has no root in F .1 shows all the subﬁelds of the extension Q( 3 2. Set Roots(p. the set of roots of p(X) in F . ζ3 )/Q. ζ3 )
Q
√ Figure 3. F ) denote the set of all monomorphisms L −→ F which ﬁx the elements of K.1. 3 2 ζ3 lying in the extension √ √ √ Q( 3 2.33. √ Figure 3. F ) = AutK (F ) since every injective Klinear transformation is surjective and so invertible. ζ3 ) √ with [Q( 3 2.32.
C
2
R
∞
VV ii r iiii rr V 2 iiii rrrr i 2 VVV iiii rr V ii r √ 2 VV 3 √ iiii √ Q( 3 2 ζ3 ) VVV Q( 3 2) Q( 3 2 ζ3 ) `` vvv VV `` vvv VV `` vvv VV `` vvv `` 3 vv3 vvv ` 3 Q(ζ3 ) vvv ``` ss vvv `` 2 ss s vvv `` sss ` vvv ss s
2
∞
√ Q( 3 2. then we do have MonoK (F.√ √ 2 and the second factor has the nonreal complex roots 3 2 ζ3 . ζ3 ) : Q] = 6. 3. We could also have begun by adjoining ζ3 to form the extension Q(ζ3 )/Q. F ) and MonoK (F. but none of √ the roots of p(X) lie in this ﬁeld so the extension Q( 3 2. ζ3 )/Q(ζ3 ) of degree 3 is obtained by adjoining one and hence all of the roots. let MonoK (L. We always have AutK (F ) ⊆ MonoK (F. Monomorphisms between extensions 3. So the splitting subﬁeld of X 3 − 2 in C over Q is Q( 3 2. Let F/K be an extension and p(X) ∈ K[X]. Definition. The subﬁelds of Q( 3 2. Remark. This is always a ﬁnite set which may of course be empty. ζ3 )/Q( 3 2) of degree 2.
These mappings have images √ √ 3 3 α0 Q( 2) = Q( 2). t0
φt =εt ◦ε−1 t
0
K(t0 ) o
εt0 ∼ =
K[X]/(p(X))
εt
/* F
Notice that if φ : K[X]/(p(X)) −→ F is any homomorphism extending the identity function on K. Q( 3 2)) contains only the identity √ since Q( 3 2) R. C) = MonoQ (Q( 2).
43
√ √ 2 3 3 α2 Q( 2) = Q( 2 ζ3 ).
. hence every such homomorphism arises this way. Of course.36. giving monomorphisms id. Then if t ∈ F is a root of p(X). Here minpolyQ. then the coset X + (p(X)) must be sent by φ to a root of p(X) in F . Hence = √ √ the Qmonomorphisms we want send 2 to ± 2 which are the complex roots of X 2 − 2. C) has 3 elements but MonoQ (Q( 3 2). We will see that this is not always true. √ We can replace C by Q( 2) to obtain √ √ √ √ MonoQ (Q( 2). Q( 3 2)) contains only the identity function. Example. α1 . 3. √ √ √ 2 Each of the above roots corresponds to one of the subﬁelds Q( 3 2). √ 3. α2 : Q( 3 2) −→ C given by √ √ √ √ 3 3 3 3 α0 (a + b 2 + c( 2)2 ) = a + b 2 + c( 2)2 . C) has two elements. α : Q( 2) −→ C. MonoQ (Q( 3 2). Example. and let F/K be an extension. If we ﬁx one such root t0 and identify K[X]/(p(X)) with K(t0 ) via εt0 . We have Q( 2) ∼ Q[X]/(X 2 − 2) where X 2 − 2 is irreducible over Q. Q( 3 2 ζ3 ) or Q( 3 2 ζ3 ) √ of C and there are 3 monomorphisms α0 . Proposition. then each root of p(X) in F gives rise to a monomorphism φt = εt ◦ ε−1 : K(t0 ) −→ F for which φt (t0 ) = t. √ √ √ 3. Show that MonoQ (Q( 2). F ) given by t ←→ φt . F ) ←→ MonoK (K(t0 ). As two of these roots are not real. √ √ √ √ 3 3 3 3 2 α1 (a + b 2 + c( 2)2 ) = a + b 2 ζ3 + c( 2)2 ζ3 .35. √ √ 3 3 α1 Q( 2) = Q( 2 ζ3 ). Q( 2)) = AutQ (Q( 2)). where φt : K(t0 ) −→ F has the eﬀect φt (t0 ) = t. √ Solution. Show that MonoQ (Q( 3 2). Let p(X) ∈ K[X] be an irreducible polynomial with t0 ∈ F be a root of p(X). 3 2 ζ3 . there is one such monomorphism for each root of p(X) in F . √ √ Solution. This discussion is summarized in the following result. √ √ √ 2 √ 3 3 3 3 α2 (a + b 2 + c( 2)2 ) = a + b 2 ζ3 + c( 2)2 ζ3 . In √ fact both possibilities occur. Then there is a bijection Roots(p.Suppose that p(X) ∈ K[X] is an irreducible polynomial which we might as well assume is monic. 3 √ 2 √ √ 3 2 ζ3 . √2 (X) = X 3 − 2 and there are 3 complex roots 3 2.34. Let F/K be a ﬁeld extension. the evaluation homomorphism εt : K[X] −→ F factors through the quotient monomorphism εt : K[X]/(p(X)) −→ F whose image is K(t) F . where √ √ α(a + b 2) = a − b 2.
We build up the list of monomorphisms in stages. There are again two possibilities (√ (√ √ 2) √ 2) 3 3 2 −→ 3 2 ζ3 2 −→ 3 2 ζ3 ′ . 2 ζ3 −→ ζ3 √ √ Next we consider monomorphisms that send 3 2 to 3 2 ζ3 . It is also worth remarking that  AutQ (Q( 3 2. each monomorphism α ∈ MonoK (L. In the case of ﬁnite. αp : Roots(p. F ) restricts to a function αp : Roots(p. so α maps Roots(p. (ii) If α ∈ MonoK (L. F ). extensions it is the key to understanding the automorphism group and this is a fundamental insight of Galois Theory. Proof. (ii) From (i). The possibilities are ) (√ ) (√ √ √ 3 3 2 −→ 3 2 ζ3 2 −→ 3 2 ζ3 ′ . where α0 has the eﬀect (√ √ ) 3 2 −→ 3 2 α0 : . C). L) −→ Roots(p. ζ3 ). α2 : . Example. C) = MonoQ (Q( 2. L) we have p(α(u)) = α(p(u)) = α(0) = 0. √ so these form a group. ζ3 ) = Q( 3 2)(ζ3 ) and that ζ3 is a root of the irreducible cyclotomic √ polynomial Φ3 (X) = X 2 + X + 1 ∈ Q( 3 2)[X]. F ) which is an injection. These form the subset √ √ 3 3 MonoQ( √2) (Q( 2. Part (ii) says that any automorphism of L/K permutes the set of roots in L of a polynomial p(X) ∈ K[X]. √ 3. ζ3 ). ζ3 ). 3 √ √ We know that Q( 3 2. L) −→ Roots(p. or more generally algebraic.38. It is a nice exercise to show that AutQ (Q( 3 2. α1 : . ζ3 ) : Q]. √ √ First consider monomorphisms that ﬁx 3 2 and hence ﬁx the subﬁeld Q( 3 2). L). So there are two monomorphisms id. α1 : 2 ζ3 −→ ζ3 ζ3 −→ ζ3 √ √ 2 Finally we consider monomorphisms that send 3 2 to 3 2 ζ3 . hence it is also surjective by the Pigeon Hole Principle. √ Solution. ζ3 )) = AutQ (Q( 2. ζ3 ). (i) For u ∈ Roots(p. ζ3 )) ∼ S3 . L) into Roots(p. ζ3 )) since again we can send ζ3 to either ζ3 2 or ζ3 . ζ3 )) = [Q( 3 2. This time we have 2 distinct ways √ √ to extend to elements of MonoQ (Q( 3 2. Determine MonoQ (Q( 3 2. (i) For p(X) ∈ K[X].
44
. α0 ﬁxing √ Q( 3 2).3. ζ3 ). We have already met the extension Q( 3 2. then αp : Roots(p.30 and we will make use of information from there. Q( 3 2.37. Proposition. L) −→ Roots(p. C). Since α is an injection its restriction to Roots(p. ζ3 ). L) is a bijection. Q( 2. α2 : 2 ζ3 −→ ζ3 ζ3 −→ ζ3 These are all 6 of the required monomorphisms. Thus αp : Roots(p. L) −→ Roots(p. C) ⊆ MonoQ (Q( 2. ζ3 )/Q in Example 3. L) is an injective function from a ﬁnite set to itself. Let F/K and L/K be extensions. the symmetric = √ √ group on 3 objects. ζ3 )). L) is a bijection. L) ⊆ L is also an injection. This gives us a strong hold on the possible automorphisms. It is also the case here that √ √ √ √ 3 3 3 3 MonoQ (Q( 2.
Suppose that u ∈ Lalg . . Theorem (Fundamental Theorem of Algebra for C). Then αq : Roots(q. An extension F/K is called an algebraic closure of K if F is algebraic over K and algebraically closed. L) −→ Roots(q. 3. Is there an algebraically closed ﬁeld F containing K? By taking F alg we might as well ask that such a ﬁeld be algebraic over K. We must show that αalg is a bijection by showing it is surjective.4. K } eee ee }} ee }} e }} ~} φ / F2 F1 Hence algebraic closures are essentially unique. This shows that v ∈ im α and so αalg is surjective. . ud ∈ C and this is unique apart from the order of the roots uj . 3. (i) There is an algebraic closure of K. 3. Now Roots(q. Every nonconstant polynomial p(X) ∈ C[X] has a factorization p(X) = c(X − u1 ) · · · (X − ud ). hence v = αq (w) = α(w) for some w ∈ Roots(q. and it is also ﬁnite. Let K be a ﬁeld. . 3. Algebraic closures An important property of the complex numbers is that C is algebraically closed.
45
.37(ii). 3. u1 . Let K be a ﬁeld. (ii) Let F1 and F2 be algebraic closures of K. Let K be a ﬁeld.44. where c. See [3] for a proof using Zorn’s Lemma (see Axiom 3. Definition.43. L) ̸= ∅ since it contains v. It is natural to pose the following question. Then α restricts to an automorphism αalg : Lalg −→ Lalg .41.48) which is logically equivalent to the Axiom of Choice. Then p(α(u)) = α(p(u)) = α(0) = 0. Question.42. L). 3. Theorem.40. Corollary. so α maps Lalg ⊆ L into itself and therefore gives rise to a restriction αalg : Lalg −→ Lalg which is also a monomorphism. Every nonconstant polynomial p(X) ∈ C[X] has a root in C. Proof. . L) ⊆ Lalg .We end this section with another useful result.39. Let L/K be an extension and α ∈ MonoK (L. 3. Let v ∈ Lalg and suppose that q(v) = 0 for some q(X) ∈ K[X] of positive degree. say p(u) = 0 for some p(X) ∈ K[X] of positive degree. Proposition. Then there is an isomorphism φ : F1 −→ F2 which ﬁxes the elements of K. L) is a bijection by Proposition 3. Proof.
) consists of a set X and a binary relation such that whenever x.
3. There are some immediate consequences of Theorem 3. Definition.44(ii). . We will temporarily write . the Monomorphism Extension Theorem. Definition.16. Let (X.49. every element of L is algebraic over K.44. all of its roots are in fact algebraic over K since (L)alg is algebraic over K.45. (ii) Every nonconstant polynomial in (L)alg [X] has a root in L. y ∈ X. Proposition.
46
. y. x then x = y. Hence these roots lie in (L)alg . 3. Axiom (Zorn’s Lemma). y or y x is true.Because of the uniqueness we usually ﬁx some choice of algebraic closure of K and write K or K alg cl . which we will ﬁnd useful. 3.
For example. (i) By Proposition 3. indeed.
x
y
=⇒
y = x. Suppose that φ0 : L −→ K is a monomorphism ﬁxing the elements of K. by Proposition 3. ) is totally ordered if for every pair x. Then X has a maximal element. at least one of x
3.
(X. • y ∈ X is an upper bound for Y if for every y ∈ Y . 3. • x x. We are already familiar with the example C = C. (i) If L/K is an algebraic extension. Proof.47. Since L is algebraically closed it is an algebraic closure of K. then L = K. which shows that it is algebraically closed. y • An element x ∈ X is a maximal element of X if y. A partially ordered set (X. (ii) If L/K is an extension. ) be a partially ordered set in which every totally ordered subset has an upper bound. Again the proof uses Zorn’s Lemma which we state below. ) be a partially ordered set and Y ⊆ X.16. First we need some deﬁnitions. referring to it as the algebraic closure of K. Theorem (Monomorphism Extension Theorem). Let K be a ﬁeld. z ∈ X.46. There is a stronger result than Theorem 3. Let (X. E1 = E2 to indicate that for extensions E1 /K and E2 /K there is an isomorphism E1 −→ E2 ﬁxing the elements of K. . we have Q = Calg and R = C. Let M/K be an algebraic extension and L/K M/K.48. • if x y and y • if x y and y
z then x z. then so is L/K and (L)alg = K.
We order X using the relation for which (F1 .Then there is an extension of φ0 to a monomorphism φ : M −→ K.
8K ¡@ ¡¡ ¡¡ ¡¡ ¡ ¡¡ M ¡¡ φ0 ¡¡ ¡¡ ¡¡
φ
L
K
=
/K
Proof. Let u.50. Also there is a function θ : F −→ K deﬁned by θ(u) = θ(u)
whenever u ∈ F for (F. This factors through the quotient ring M0 [X]/(minpolyM0 .u (X) ∈ K[X]. θ0 ) and ′′ (M0 .
47
. 3. Then for every (F. Example. θ) ∈ Y . Then (X. θ) (F . Let ∪ F = F. But then (M0 . If minpolyM0 . v ∈ K say. so (F . θ0 ). θ1 ) (F2 . θ0 ).51.
(F. Let u ∈ K and suppose that p(X) = minpolyK. contradicting the maximality of (M0 . θ). θ0 ). θ) ∈ Y we have (F.22 of the polynomial ring M0 [X] applied to ′ the monomorphism θ0 : M0 −→ K yields a homomorphism θ0 : M0 [X] −→ K extending θ0 ′ and for which θ0 (u) = v. θ′ ) ∈ Y then θ′ (u) = θ(u). Definition. θ0 ) ̸= (M0 (u). there is a monomorphism φv : K(u) −→ K with φv (u) = v. Suppose that M0 ̸= M . Since M is algebraic / over K it is also algebraic over M0 . The Homomorphism Extension Property 1. θ2 ) whenever F1 F2 and θ2 extends θ1 . θ) is an upper bound for Y . then the polynomial f (X) = θ0 (a0 ) + · · · + θ0 (an−1 )X n−1 + X n ∈ (θ0 M0 )[X] is also irreducible and so it has a root v in K (which is also an algebraic closure of θ0 M0 K). 3. We consider the set X consisting of all pairs (F. so there is an element u ∈ M for which u ∈ M0 . so θ is welldeﬁned. θ0 ) (M0 (u). θ). By Zorn’s Lemma there must be a maximal element of X. (M0 . where F/L M/L and θ : F −→ K extends φ0 . ) is a partially ordered set.u (X) = a0 + · · · + an−1 X n−1 + X n . It is straightforward to check that if u ∈ F ′ for (F ′ . Then v is conjugate to u over K or is a conjugate of u over K if there is a monomorphism φ : K −→ K ﬁxing K for which v = φ(u).θ)∈Y
Then F /L
M/L. Then for any other root of p(X). hence u is algebraic over M0 . Suppose that Y ⊆ X is a totally ordered subset. Hence M0 = M and so we can take φ = θ0 . This extends to a monomorphism φ : K −→ K.u (X)) to ′′ ′′ give a monomorphism θ0 : M0 (u) −→ K extending θ0 . v ∈ K.
So ∂(f (X)) = 0 if and only if f (X) = a0 ∈ K. i.
48
. It is also clear that every polynomial g(X) ∈ K[X] has the form g(X) = ∂(f (X) where f (X) is an antiderivative of g(X). then ker ∂ = K and ∂ is surjective.52.. v ∈ K.50. K). If f1 (u) = 0 then u is a multiple or repeated root of f (X). the function ∂ : K[X] −→ K[X] given by ∂(f (X)) = f ′ (X) = a1 + 2a2 X + · · · + dad X d−1 .u (v) = 0. If minpolyK.54. i. (iii) If char K = 0.e. Multiplicity of roots and separability Let K be a ﬁeld. We need to understand more clearly when an irreducible polynomial has a multiple root since this turns out to be important in what follows. Suppose that f (X) ∈ K[X] and u ∈ K is a root of f (X). for f (X). where f (X) = a0 + a1 X + a2 X 2 + · · · + ad X d with aj ∈ K. then a0 + a1 u + · · · + ad−1 ud−1 + ud = 0 and so a0 + a1 v + · · · + ad−1 v d−1 + v d = φ(a0 + a1 u + · · · + ad−1 ud−1 + ud ) = 0.u (X) = a0 + a1 X + · · · + ad−1 X d−1 + X d . 3.e. 3. s 0.53. If u. f (u) = 0. Suppose that v = φ(u) for some φ ∈ MonoK (K.e. g(X) ∈ K[X]. Proof. then ker ∂ = {h(X p ) : h(X) ∈ K[X]} and im ∂ is spanned by the monomials X k with p (k + 1).. Lemma. If f1 (u) ̸= 0 then u is a simple root of f (X). i. then v is conjugate to u over K if and only if minpolyK. Consider the formal derivative on K[X]. ∂(f (X)g(X)) = ∂(f (X))g(X) + f (X)∂(g(X)). Proof. 3. (ii) ∂ is a derivation. The formal derivative ∂ : K[X] −→ K[X] has the following properties. it suﬃces to verify this for the case where f (X) = X r and g(X) = X s with r. Then we can factor f (X) as f (X) = (X − u)f1 (X) for some f1 (X) ∈ K[X]. (i) This is routine.3. (i) ∂ is Klinear.5. But then ∂(X r+s ) = (r + s)X r+s−1 = rX r−1 X s + sX r X s−1 = ∂(X r )X s + X r ∂(X s ). (iii) If f (X) = a0 + a1 X + a2 X 2 + · · · + ad X d then ∂(f (X)) = 0 ⇐⇒ a1 = 2a2 = · · · = dad = 0. Proposition.. (ii) By Klinearity. Definition. The converse follows from Example 3. (iv) If char K = p > 0.
If f (X) is irreducible in K[X] then a root u is a multiple root if and only if f ′ (X) = 0.58.
Also. We now apply the formal derivative to detect multiple roots. Solution.
. If char K = 0 and f (X) is irreducible in K[X]. Hence 2i is a multiple root of f (X). Using Long Division and the Euclidean Algorithm we ﬁnd that gcd(f (X). .60. then the multiplicity of u in p(X) is the maximum m such that p(X) = (X − u)m q(X) for some q(X) ∈ L[X]. Example.12) and which vanishes at u. Show that 2i is a multiple root of f (X) = X 4 + 8X 2 + 16. X 4 + 8X 2 + 16 = (X 2 + 4)2 . Let p > 0 be a prime and suppose that L/Fp is an extension. 3. An irreducible polynomial p(X) ∈ K[X] is separable over K if every root of p(X) in an extension L/K is simple. 3. Let f (X) ∈ K[X] have a root u ∈ L for some extension L/K. 3. im ∂ is spanned by the monomials X k for which ∂(X k+1 ) ̸= 0.57. Let K be a ﬁeld and let K be an algebraic closure. Example.(iv) For a monomial X m . 3. We have f ′ (X) = 4X 3 +16X. which are the ones with p (k + 1). Hence in K[X]. Example. Show that each of the roots of f (X) = X p − 1 in L is multiple. 3.55. p(X) = c(X − u1 )m · · · (X − uk )m . We have f ′ (X) = ∂(X p − 1) = pX p−1 = 0. where 2i is also a root of X 2 + 4. this can only happen if char K > 0. so for any root ζ of f (X). Working in L[X]. Then u is a multiple root of f (X) if and only if f (X) and f ′ (X) have a common factor of positive degree in K[X] which vanishes at u. In fact.61. Solution. Later we will see that 1 is the only root of X p − 1. Corollary.62.
49
Solution. so if ζ is any root of f (X) then f ′ (ζ) = 0. In particular. this is equivalent to requiring that p′ (X) ̸= 0. 3. Hence u is a multiple root if and only if f (X) and f ′ (X) have a common factor in L[X] (and hence in K[X] by Proposition 3.56. f ′ (X)) = X 2 + 4. then every root of f (X) is simple. let f (X) = (X − u)f1 (X). For n 1. f ′ (ζ) = nζ n−1 ̸= 0. Corollary. Using this we see that ∂(a0 + a1 X + a2 X 2 + · · · + ad X d ) = 0 ⇐⇒ am = 0 whenever p m. uk ∈ K. show that each of the roots of f (X) = X n − 1 in C is simple. 3. ∂(X m ) = mX m−1 and this is zero if and only if p  m.
so f ′ (u) = f1 (u). . . Proof. By Corollary 3. where c ∈ K and m 1. Definition.59. . then the multiplicities of the uj are equal. Proposition. 3. If u ∈ L is a multiple root of p(X). Then
′ f ′ (X) = f1 (X) + (X − u)f1 (X). Proposition. We have f ′ (X) = ∂(X n − 1) = nX n−1 . If the irreducible polynomial p(X) ∈ K[X] has distinct roots u1 . so this is obvious.56.
34. .
50
. . Any ﬁnite extension L/K can be built up from a succession of simple extensions (3. Corollary. By Proposition 3.65. Corollary.1) K(u1 )/K. uk−1 ).67. then [K(u) : K] = (K(u) : K). An algebraic extension L/K of a ﬁeld of characteristic 0 is separable by Corollary 3. 3. An algebraic element u ∈ L in an extension L/K is separable if its minimal polynomial minpolyK. .34. there is a monomorphism φv : K(u) −→ K for which φv (u) = v. . . Definition. p(X) = c(X − u1 ) · · · (X − uk ). . the coeﬃcients of p(X) are ﬁxed by φv .u . L = K(u1 . For a ﬁnite simple extension K(u)/K. . Example. Definition. Let K be a ﬁeld and let u ∈ K. and therefore the multiplicity of v must be at least m. Let K be a ﬁeld and let K be an algebraic closure. where p1 (X) ∈ K[X] is obtained applying φv to the coeﬃcients of p1 (X).u (X).68. Proof.u (X) . u2 )/K(u1 ). . This follows from Proposition 3. Then the number of distinct conjugates of u is deg minpolyK. Definition. If K(u)/K is separable.63. When p(X) is viewed as an element of K(u)[X]. K(u1 . . · · · . Let L/K be a ﬁnite extension. K). . Then φv ((X − u)m p1 (X)) = (X − u)m p1 (X).u (X) ∈ K[X] is separable. If the irreducible polynomial p(X) ∈ K[X] has distinct roots u1 . 3. K). . . 3.
So we can use the following to compute (L : K) = (K(u1 .64. uk )/K(u1 .34 applied to the case L = K. Let u ∈ K be a root of p(X) and suppose that it has multiplicity m.69. . An algebraic extension L/K is called separable if every element of L is separable over K. 3. m where m is the multiplicity of u in minpolyK.Proof.57.66. o 3. . The separable degree of L over K is (L : K) =  MonoK (L. uk ) : K). (K(u) : K) =  Roots(minpolyK. 3. Now by Corollary 1. 3. so we can write p(X) = (X − u)m p1 (X) where p1 (X) ∈ K(u)[X] and p1 (u) ̸= 0. and so (X − v)m p1 (X) = (X − u)m p1 (X). . uk ∈ K which are all simple then in K[X]. . Interchanging the rˆles of u and v we ﬁnd that the multiplicities of u and v are in fact equal. Lemma. Now let v ∈ K be any other root of p(X). (X − v)m must divide p1 (X) in K[X]. where c ∈ K and k = deg p(X).
. suppose that (L : K) = [L : K]. hence β −1 ◦ β ′ ∈ MonoL (M. choose any extension to a monomorphism β : K −→ K. . .u (X) is separable. K). . K) agrees with the number of extensions of β to a monomorphism M −→ K. u2 ) : K(u1 )] · · · [K(u1 . 3. . u is separable. K) let αL ∈ MonoK (L. uk−1 )). . Proposition. uk ) : K(u1 . Proof. So given β ∈ MonoK (L.u (X). We must show that for each u ∈ L.u (X)  minpolyK.62 and 3. . Then (M : K) = (M : L)(L : K). . . .3. we need to show that the number of such α is always (M : L) =  MonoL (M. Expressing L/K in terms of a sequence of simple extensions as in (3. Let L/K be a ﬁnite extension. In eﬀect. Since (L : K) =  MonoK (L. . . uk−1 )]. . then β −1 ◦ β ′ (u) = β −1 (β(u)) = u.69 to each of these intermediate separable simple extensions to obtain (L : K) = [L : K]. we have (L : K) = (K(u1 ) : K) · · · (L : K(u1 . . K).39. K) has the form β = αL for some α ∈ MonoK (M. For α ∈ MonoK (M. Clearly E/K is also separable. . Corollary.72. . and therefore L/E is separable. K). K) be its restriction to L. . each element of MonoK (L. Proof. so is minpolyE.u (X). .34 we know that this is true. uk ) : K(u1 .71. Proof. and in the polynomial ring E[X] we have minpolyE. .49. We have (L : K) = (L : E) (E : K) and [L : K] = [L : E] [E : K]. uk−1 )) divides [K(u1 . so the desired result follows. uk−1 )].6(ii) which gives [L : K] = [K(u1 ) : K] [K(u1 . For each k. The general result follows by building up L/K as a sequence of simple extensions as in (3. K) gives rise to a monomorphism β ◦ γ : M −→ K which extends β. so (M : L) =  MonoL (M. . For the extensions K(u)/K and L/K(u) we have (L : K) = (L : K(u)) (K(u) : K)
51
. .70. If L/K is a simple extension then by Propositions 3. Therefore we have the desired formula (M : K) = (M : L)(L : K). As minpolyK. notice that if u ∈ L.1). . then for any u ∈ L. . 3. Now we can apply Lemma 3. . . . uk ) : K(u1 . this shows that there is a bijection { } extensions of β to monomorphism a M −→ K ←→ MonoL (M. Conversely. u is algebraic over E. K). by Proposition 3. Then (L : K)  [L : K]. K). . By the Monomorphism Extension Theorem 3. Let L/K and M/L be ﬁnite extensions. . Of course. . so to verify that (L : K) = [L : K] it suﬃces to show that (L : E) = [L : E] and (E : K) = [E : K]. restricting to M K we obtain a monomorphism M −→ K. . Then L/K is separable if and only if (L : K) = [L : K]. . Proposition. Now for any extension β ′ : M −→ K of β we can form the composition β −1 ◦ β ′ : M −→ K. β is an automorphism. K). Suppose that L/K is separable. If K E L. so every element β ∈ MonoK (L. uk−1 )]. each γ ∈ MonoL (M. For the converse. [L : K] = [K(u1 ) : K] · · · [L : K(u1 .1) and then using Theorem 2. Let L/K be a ﬁnite extension. (K(u1 . K) extends to a monomorphism M −→ K. .
. v). 3. we can choose an element t ∈ K for which t ̸= u − ui vj − v
whenever j ̸= 1. .73. Definition.75. but h(vj ) ̸= p(ui ) = 0 for any j ̸= 1 by construction of t. The case where K is a ﬁnite ﬁeld will be dealt with in Proposition 5.and [L : K] = [L : K(u)] [K(u) : K]. vs . So we will assume that K is inﬁnite. Deﬁne the polynomial (of degree r) h(X) = p(w − tX) ∈ K(w)[X] ⊆ L[X]. s for which [L : K(u)] = r(L : K(u)) and [K(u) : K] = s(K(u) : K). Then L has a primitive element. For a ﬁnite simple extension L/K. Therefore M/K is separable.72 this implies that L/K and M/L are separable. By Corollary 3. q(X) ∈ K[X] be the minimal polynomials of u and v over K.
52
. namely v. [M : L][L : K] = (M : L)(L : K). hence [M : K] = [M : L][L : K] = (M : L)(L : K) = (M : K). Conversely. 3. which can only happen if r = s = 1. . .16. since (M : L) [M : L] and (L : K) [L : K]. we ﬁnd that w ̸= ui + tvj whenever j ̸= 1. hence v ∈ K(w) and so u = w − tv ∈ K(w). Let L/K and M/L be ﬁnite extensions. Let p(X). The Primitive Element Theorem 3. Thus (K(u) : K) = [K(u) : K] and so u is separable. if L/K and M/L are separable then [M : L] = (M : L) and [L : K] = (L : K). Then taking w = u + tv ∈ L. Proof. . Proof. By Proposition 3. . By the separability assumption. This can only happen if [M : L] = (M : L) and [L : K] = (L : K). there are some positive integers r. If M/K is separable then [M : K] = (M : K) and so by Proposition 3.71. 3. v). Since K is inﬁnite. while the distinct roots of q(X) are v = v1 . an element u ∈ L is called a primitive element for the extension if L = K(u). r = deg p(X) and s = deg q(X). v) K(w) and therefore K(w) = K(u.6. Proposition. ur . Then h(v) = p(u) = 0. Then M/K is separable if and only if L/K and M/L are separable. Suppose that the distinct roots of p(X) in K are u = u1 . Hence (L : K(u))(K(u) : K) = rs(L : K(u))(K(u) : K). . . q(X) ∈ K(w)[X] have exactly one common root in K. Since L is built up from a sequence of simple extensions it suﬃces to consider the case L = K(u.70. so none of the other vj is a zero of h(X). Let L/K be a ﬁnite separable extension. This shows that K(u. Theorem (Primitive Element Theorem). Now since the polynomials h(X). by separability their greatest common divisor in K(w)[X] is a linear polynomial which must be X − v.74.
√3+i (X) = 4. the polynomial (X − α1 (u)) · · · (X − αd (u)) ∈ K[X] is in K[X] and is irreducible therein.77. α1 = id. Thus we must have deg minpolyQ.
we see that minpolyQ. Let L/K be a ﬁnite extension of a ﬁeld of characteristic 0. i) : Q( 3)][Q( 3) : Q] = 2 · 2 = 4. 3. Proof.76. α3 = . = 3+1 4 ( 3 + i)( 3 − i) √ √ since ( 3 + i)−1 ∈ Q( 3 + i). √ √ There are four monomorphisms αk : Q( 3. αd are the elements of MonoK (K(u).52. √ √ √ Solution.75. i = (( 3 + i) − ( 3 − i)). Corollary.78. Proof. Since Q K. K). Proposition. i)/Q.77 we have √ √ √ √ [Q( 3. Since K(u) is separable then by Lemma 3. . i) : Q] = [Q( 3. √ √ (X 2 − 2 3X + 4)(X 2 + 2 3X + 4) = X 4 − 4X 2 + 16 ∈ Q[X]. i) = Q( 3 + i). There is a general phenomenon illustrated by Example 3. √ 3. showing that Q( 3. i) −→ Q( 3.√3+i (X) = X 4 − 4X 2 + 16. . Example. i) Q( 3 + i) and so Q( 3.√3+i (X). α4 = . i) given by (√ (√ (√ √ ) √ ) √ ) 3 −→ 3 3 −→ − 3 3 −→ − 3 . 2 2 √ √ √ √ √ are both in Q( 3 + i).3. however a ‘try it and see’ approach will often be suﬃcient. In particular. hence Now taking √ X 2 − 2 3X + 4 = minpolyQ(√3). . Hence √ √ √ 1 √ 1 √ 3 = (( 3 + i) + ( 3 − i)). Find a primitive element for the extension Q( 3.
To ﬁnd a primitive element we can always use the method suggested by the proof of Theorem 3. Then L has a primitive element.u (X) = [K(u) : K] = (K(u) : K).67 L/K is separable.√3+i (X)  (X 4 − 4X 2 + 16) in Q[X]. α2 = i −→ −i i −→ i i −→ −i
53
. d = deg minpolyK. and so minpolyQ. Notice that √ √ √ √ ( 3 − i) 1 √ ( 3 − i) −1 √ ( 3 + i) = √ = ( 3 − i) ∈ Q( 3 + i). Then working over the subﬁeld Q( 3) Q( 3. where α1 .u (X) = (X − α1 (u)) · · · (X − αd (u)). . In Example 3. i) we ﬁnd √ that i ∈ Q( 3) R and / √ √ √ √ (X − ( 3 + i))(X − ( 3 − i)) = X 2 − 2 3X + 4 ∈ Q( 3)[X]. Then minpolyK. Let u ∈ K be separable over K.77. K is inﬁnite and by Example 3. Consider 3 + i.
E is generated by some of the roots of f (X). . K) must map these roots to θ(v1 ).80. . . K).
√ √ √ √ 3 − i)(X − 3 + i)(X + 3 − i)(X + 3 + i) = X 4 − 4X 2 + 16 ∈ Q[X]. Then there is a sequence of extensions K K(u1 ) K(u1 . vk ). Let E/L and L/K be ﬁnite extensions.u1 (X) minpolyK. . 3. vk ) = E. indeed these are known as Galois extensions. Now suppose that E is a splitting ﬁeld for g(X) ∈ K[X]. Hence E is a splitting ﬁeld for f (X) over K. .7. . vk ) = K(θ(v1 ). . . . . then E is the splitting ﬁeld of f (X) over L. φE = E if and only if φE E. . Proof. . Since θ permutes the roots vj .u2 (X) · · · minpolyK. . Proof. . If E/K is a normal extension then whenever an irreducible polynomial p(X) ∈ K[X] has a root in E. Also. f (X) splits in E. . . vk are the distinct roots of g(X) in E. un ) = E
Construct a polynomial by taking f (X) = minpolyK.81. then by Remark 3. θ(vk ) which are also roots of g(X) and therefore lie in E (see Proposition 3. . E/K is normal if φE = E for every φ ∈ MonoK (E. Corollary. (X − 3. so √ √ α3 ( 3 + i) = (− 3 + i). .82. So we have [Q( 3 + i) : Q] = 4 and Q( 3 + i) = Q( 3. Suppose that E/K is normal. K). . so that E = K(v1 . 3. A ﬁnite extension E/K is normal if and only if it is a splitting ﬁeld over K for some polynomial f (X) ∈ K[X]. . . .52 each pair of roots of p(X) is conjugate over K and one can be mapped to the other by a monomorphism K −→ K which must map E into itself. If E/K is normal then E/L is normal. √ √ √ Hence this polynomial is irreducible. In Chapter 4 we will see that separable normal extensions play a central rˆle in Galois Theory. . 3. u2 ) ··· K(u1 .un (X). Remark.80. √ √ α4 ( 3 + i) = (− 3 − i). o
54
. where v1 . Definition. . . Normal extensions and splitting ﬁelds Let K be an algebraic closure for the ﬁeld K and let E/K K/K be a ﬁnite extension. i). . it splits in E since by Lemma 3.34). Theorem. . . we have θE = θK(v1 . θ(vk )) = K(v1 . These result makes it easy to recognize a normal extension since it is suﬃcient to describe it as a splitting ﬁeld for some polynomial over K. 3.79. Then by Remark 3.Then √ √ α2 ( 3 + i) = ( 3 − i).32. If φ ∈ MonoK (E. Now any monomorphism θ ∈ MonoK (E. If E is the splitting ﬁeld of a polynomial f (X) ∈ K[X] over K.
in each case ﬁnding it minimal polynomial over Q.
p2 (X) = X 6 −2.] 3.. Q( 2. i)/Q.] √ 3. p4 (X) = X 4 +5X 3 +10X 2 +10X+5. How does g(X) factor in k(T )[X]? √ √ √ √ 3. Let k be a ﬁeld of characteristic char k = p > 0 and k(T ) be the ﬁeld of rational functions in T over k.20: Let L/K be a ﬁnite extension.4.3. as claimed in the = solution of Example 3. i. un ).Exercises on Chapter 3 3. L = K(w) for some w ∈ L. the symmetric group on 3 elements. If there are only ﬁnitely many subextensions F/K then L/K is simple.1. Prove Proposition 3. Show that Q(u)/Q is not a normal extension. .] 3.e. consider p4 (Y − 1) ∈ Q[Y ]. v). Finding splitting subﬁelds E polynomials.38. . i)/Q. then use induction on n to prove the general case L = K(u1 . [Hint: work out the eﬀect of each automorphism on the three roots of the polynomial X 3 − 2.
[Hint: First deal with the case where L = K(u. Q( 3.e.75.5. of degree 2) extension E/K is normal. .. ζ3 )) ∼ S3 . Is such an extension always separable? 3. 10)/Q. √ Q( 4 3. Prove that AutQ (Q( 3 2.2. Show that the polynomial g(X) = X p − T ∈ k(T )[X] is irreducible and has a multiple root in k(T ).
[Hint: for p4 (X). 3. C over Q and determine [E : Q] for each of the following p3 (X) = X 4 +2.
55
. Prove the following converse of Proposition 3. Let f (X) ∈ Q[X] be an irreducible polynomial of odd degree greater than 1 and having only one real root u ∈ R. i)/Q. Find primitive elements for the extensions Q( 5. Let K be a ﬁeld.7.] 3. .6. p1 (X) = X 4 −X 2 +1.8. L/K. [Hint: look for elements of high degree over Q. or use the method of proof of Theorem 3.2. Show that every quadratic (i.
.
Definition. [E : K] = (E : K) and also every monomorphism φ ∈ MonoK (E. the group Gal(E/K) = AutK (E) is called the Galois group of the extension or the Galois group of E over K. let K be a ﬁeld. in particular we will explain how these are related through the Galois Correspondence.1. Galois extensions 4.49. Let E/K a ﬁnite Galois extension and u. we also say that v is a conjugate of u.1) implies (4.1. K) maps E into itself. Notice that Equation (4.2. Then v is conjugate to u if there is a φ ∈ Gal(E/K) for which v = φ(u).3.1)  AutK (E) = (E : K) = [E : K]. v ∈ E. From Section 3.5 we know that for such a Galois extension E/K.
4.
We can also reformulate the notion of conjugacy introduced in Deﬁnition 3.
>K ~~ ~ ~~ ~~ ~ ∼ = E φ / E
φ
E
K
=
/K
Also.51. A ﬁnite extension E/K is a (ﬁnite) Galois extension if it is both normal and separable. by the Monomorphism Extension Theorem 3. Throughout the chapter. hence restricts to an automorphism of E which will be denoted φE . The elements of Gal(E/K) are called (Galois) automorphisms of E/K. 4. For a ﬁnite Galois extension E/K. Definition. K) ←→ AutK (E) and we have (4.
57
. every automorphism α ∈ AutK (E) extends to a monomorphism E −→ K ﬁxing elements of K. So there is a bijection MonoK (E.CHAPTER 4
Galois extensions and the Galois Correspondence
In this Chapter we will study the structure of Galois extensions and their associated Galois groups.2)  Gal(E/K) = (E : K) = [E : K]. Definition. 4.
We have √ √ √ √ √ √ [Q( 2. Solution. Example. We also know that E is a simple extension of K since E/K is separable. φ −→ φE F
is a surjective group homomorphism.6. Here is a slightly diﬀerent way to understand this. Working with Galois groups
F θ −→ θE . 3)/Q) as a subgroup of the group of permutations of the roots of minpolyQ.It is easy to see that for u.u (X) ∈ Q[X]. It is often convenient to use these facts to interpret elements of the Galois group as permutations of the roots of some polynomial which splits over E.
4.2. Example. 3. The proof of the next result is left as an exercise. − 3 and numbering them from 1 to 4. − 2. E 4. then the function AutK (K. √ √ 4. 3)/Q) as a subgroup of the group of permutations of the roots of (X 2 − 2)(X 2 − 3) ∈ Q[X]. Furthermore. α4 ←→ (1 2)(3 4). K) −→ AutK (E.
Let E/K be a ﬁnite Galois extension. If F/K K/K is any ﬁnite normal extension with E then there is a surjective group homomorphism AutK (F. 3) : Q] = [Q( 2. First notice that every element φ ∈ AutK (K. Hence E is a splitting ﬁeld for the minimal polynomial of any primitive element for E/K. F ) −→ AutK (E. Using a primitive element u for the extension. If E/K is a ﬁnite Galois extension. and the following nontrivial elements of the Galois group together with the element identity α1 = id: √ √ √ √ √ √ 2 −→ − 2 2 −→ 2 2 −→ − 2 √ √ √ √ √ √ − 2 −→ − 2 −→ − 2 − 2 −→ 2 2 √ √ √ √ . E). α4 = √ . α3 = √ . if F/K is any ﬁnite normal extension with E F . this minimal polynomial has degree [E : K].5. hence gives rise to an automorphism φE : E −→ E. v ∈ K. α2 = 3 −→ 3 3 −→ − 3 3 −→ − 3 √ √ √ √ √ √ − 3 −→ − 3 − 3 −→ − 3 −→ 3 3 √ √ √ √ Writing the roots in the list 2. 4. Proposition.
58
. K) we have (φF )F = φE . 3) : Q( 2)] [Q( 2) : Q] = 4. Then we know that E is a splitting ﬁeld for some polynomial over K since E/K is normal. K) restricts to a monomorphism E −→ K whose image is contained in E. every automorphism F θ : F −→ F restricts to an automorphism θE : E −→ E. there is a ﬁnite Galois extension E/K in which v is a conjugate of u if and only v is a conjugate of u over K in the old sense. Similarly. Describe the Galois group Gal(Q( 2. describe the Galois group √ √ Gal(Q( 2. E). for φ ∈ AutK (K. α3 ←→ (3 4). these automorphisms correspond to the following permutations in S4 expressed in cycle notation: α2 ←→ (1 2).4.
6. . and after numbering them accordingly. . 2 2 √ √ √ we easily ﬁnd that 2. i) = Q(ζ8 ). Recollection. y ∈ X. (iii)  Gal(E/K) divides n! and is divisible by n.5 and 4. (ii) Gal(E/K) can be identiﬁed with a subgroup of the group of permutations of Roots(f. Since [Q( 2. Then the following are true. (1 4)(2 3)} Noticing that 1 1 ζ8 = √ + √ i. Example. we have √ √ Q( 2.
Next we summarize the properties of Galois groups that can be deduced from what we have established so far. The Galois extension Q(ζ8 )/Q has degree [Q(ζ8 ) : Q] = 4 and it has the following automorphisms apart from the identity:
3 α : ζ8 −→ ζ8 .ζ (X) = Φ8 (X) = X 4 + 1
3 5 7 in the order ζ8 .
.
If we list the roots of the minimal polynomial minpolyQ. i) Q(ζ8 ). F ) denotes the set of roots of f (X) in F . the action is faithful or eﬀective if for every nonidentity element h ∈ G. If we order the roots u1 . (1 3)(2 4). ζ8 .
So the Galois group Gal(Q(ζ8 )/Q) corresponds to {id. there is an element g ∈ G such that y = gx (so there is only one orbit).
γ ←→ (1 4)(2 3). ζ8 . E). 4. i) is the splitting ﬁeld of f (X) = (X 2 − 2)(X 2 + 1) over
59
S4 . 4.8. α4 ←→ (1 4)(2 3).
β ←→ (1 3)(2 4). 2 − 3. 7 γ : ζ8 −→ ζ8 . Recall that an action of a group G on a set X is transitive if for every pair of elements x. Let E/K be a ﬁnite Galois extension. . (1 2)(3 4). ζ8 . E). We have Q( 2.7. α3 ←→ (1 2)(3 4). Theorem. Recall that for an extension F/K and a polynomial f (X) ∈ K[X]. we ﬁnd the correspondences α2 ←→ (1 3)(2 4). Listing these as √ √ √ √ √ √ √ √ 2 + 3. there is an element z ∈ X such that hz ̸= z. 5 β : ζ8 −→ ζ8 . (i) Gal(E/K) acts transitively and faithfully on Roots(f. As we have seen in Examples 4. . − 2 + 3. Notice that Q( 2. in practise it is often easier to use a not necessarily irreducible polynomial to determine and work with a Galois group. Roots(f. 3) = Q( 2 + 3) and the conjugates of u = 2 + 3 are √ √ ± 2 ± 3. Suppose that E is the splitting ﬁeld of a separable irreducible polynomial f (X) ∈ K[X] of degree n. un then Gal(E/K) can be identiﬁed with a subgroup of Sn .√ √ √ √ √ √ Solution.9. − 2 − 3. i) : Q] = 4. 4. i ∈ Q(ζ8 ). hence Q( 2. we ﬁnd that these automorphisms correspond to the following permutations in S4 :
α ←→ (1 2)(3 4).
10. v ∈ E Γ and γ ∈ Γ. 4. Then Q(f (X)) = Q(u1 .11.
γ(u−1 ) = γ(u)−1 = u−1 . For u.√ √ Q. Then Gal(Q(f (X))/Q) ∼ S3 . γ(u + v) = γ(u) + γ(v) = u + v. this ﬁxes u1 and interchanges u2 . Example. E Γ = {u ∈ E : ∀γ ∈ Γ.
Proof. While it can be hard to determine Galois groups in general. − 2. 4. Now list the roots of f (X) in the order 2. so it corresponds to the transposition (2 3). γ(uv) = γ(u)γ(v) = uv.12. special arguments can sometimes be exploited. u3 ) and in fact [Q(f (X)) : Q] = 6 since [Q(f (X)) : Q]  6 and u2 ∈ Q(u1 ) / R. we can assume that this is (1 2 3). (3 4). f (c− ) < 0. Hence Gal(Q(f (X))/Q) is isomorphic to a subgroup of S3 and so Gal(Q(f (X))/Q) ∼ S3 since the orders agree. Such examples occur when the cubic polynomial f (X) has local maximum and minimum at real values c+ and c− with f (c+ ). u2 . We also have Q(f (X)) ∩ R = Q(u1 ). It also contains an element of order 2 obtained by restricting complex conjugation to Q(f (X)). focusing on the relationship between objects of these types. = Proof. γ(u) = u}. 4. u2 . u3 . E Γ E is a subﬁeld of E containing K. f (c− ) > 0 or f (c+ ). Also. (1 2)(3 4)} S4 . f (−1) = 5. Let u1 ∈ R be the real root of f (X) and let u2 . u3 . the Galois group Gal(Q(ζ8 )/Q) = Gal(Q( 2. Subgroups of Galois groups and their ﬁxed ﬁelds Let E/K a Galois extension and suppose that Γ Gal(E/K). Lemma. u3 be the remaining complex roots. Consider the subset of elements of E ﬁxed by Γ. i −→ −i −i −→ i √ In this description. Given a Galois extension E/K. ←→ (1 2). we will next study subextensions L/K E/K and subgroups Γ Gal(E/K). Remark. 4. and observe that √ √ √ √ 2 −→ − 2 2 −→ − 2 √ √ −√2 −→ −√2 −→ 2 2 α: β: ←→ (1 2)(3 4). if u ̸= 0. i −→ −i i −→ i −i −→ i −i −→ −i √ √ 2 −→ 2 −√2 −→ −√2 γ: ←→ (3 4). This happens for example with f (X) = X 3 − 3X + 3 which has local extrema at ±1 and f (1) = 1. (1 2).3.
60
. i)/Q) corresponds to the subgroup {id. −i. = The Galois group Gal(Q(f (X))/Q) contains an element of order 3 which corresponds to a 3cycle when viewed as a permutation of the roots u1 . Suppose that f (X) = X 3 + aX 2 + bX + c ∈ Q[X] is an irreducible cubic and that f (X) has only one real root. i.
In fact we have 4. 4. Subﬁelds of Galois extensions and relative Galois groups Let E/K a Galois extension and suppose that L/K E/K (i.16. This shows that [E : E Γ ] = [E Γ (u) : E Γ ] Since Γ Gal(E/E Γ ). E Γ
EΓ. hence u ∈ E Gal(E/L) . Now each element of Gal(E/E Γ ) is also an element of Gal(E/K) and Gal(E/E Γ ) Gal(E/K). where h = Γ. we also have h = Γ  Gal(E/E Γ ) = [E : E Γ ].4. there is an automorphism θ ∈ Gal(E/L) such that θ(u) ̸= u.
Combining these two inequalities we obtain [E : E Γ ] =  Gal(E/E Γ ) = Γ = h and therefore Γ = Gal(E/E Γ ). Then E/L is also a Galois extension whose Galois group Gal(E/L) is sometimes called the relative Galois group of the pair of extensions E/K and L/K. The next result explains the connection between the two uses of the word normal which both ultimately derive from their use in Galois theory.e. we will identify its Galois group. The following is immediate. Gal(E/L) Gal(E/K). and its order is  Gal(E/L) = [E : L]. Lemma. Proposition.15. γh ..14.Finally. Clearly L E Gal(E/L) . so this is a Galois extension.
E is the ﬁxed subﬁeld of Γ. Now let the distinct elements of Γ be γ1 = id. . The relative Galois group of the pair of extensions L/K E/K is a subgroup of Gal(E/K). K L E). This shows that / Gal(E/L) Gal(E/L) = L.e. say E = E Γ (u).
61
. By Theorem 4. the extensions E/E Γ and E Γ /K are separable. Consider the polynomial of degree h f (X) = (X − u)(X − γ2 (u)) · · · (X − γh (u)) ∈ E[X]. so Lagrange’s Theorem implies that Γ divides  Gal(E/E Γ ).73.75 it is simple. .
By Proposition 3. We know that E/E Γ is separable. i. E L and therefore E We need to understand when Gal(E/L) Gal(E/K) is actually a normal subgroup. Definition.
Proof. . if t ∈ K then γ(t) = t. Let L/K E/K. Notice that by deﬁnition Γ Gal(E/E Γ ). Then L = E Gal(E/L) .
Proof. h = Γ.13. Γ
[E : E Γ ] =  Gal(E/E Γ ) = Γ. .. E/E Γ is also normal. Notice that f (X) is unchanged by applying any γk to its coeﬃcients since the roots γj (u) are permuted by γk . Proposition. Suppose that u ∈ E − L. γ2 . we have Gal(E/E Γ ) = Γ and the equations [E Γ : K] =  Gal(E/K) . For Γ Gal(E/K). 4. 4. Notice that [E : E Γ ] = (E : E Γ ) =  Gal(E/E Γ ). Hence.8(i). so K 4. so by the Primitive Element Theorem 3. f (X) ∈ E Γ [X].
then for any γ ∈ Gal(E/K) and α ∈ Gal(E/L). (ii) If α ∈ Gal(E/K). then there is a group isomorphism Gal(E/K)/ Gal(E/L) − Gal(L/K). φ(v) ∈ L. (ii) If L/K is normal and hence a Galois extension. for all α ∈ Gal(E/L) and β ∈ Gal(E/K).  Gal(E/K)/ Gal(E/L) = 4. i. The Galois Correspondence and the Main Theorem of Galois Theory We are now almost ready to state our central result which describes the Galois Correspondence associated with a ﬁnite Galois extension. φ Gal(E/L)φ−1 = Gal(E/L). every monomorphism L −→ K ﬁxing K extends to a monomorphism E −→ K which must have image E. F(E/K) = the set of all subextensions L/K of E/K. Hence for every φ ∈ Gal(E/K).
62
. (i) Suppose that Gal(E/L) ▹ Gal(E/K). Now if u ∈ L.
(i) The relative Galois group Gal(E/L) of the pair of extensions L/K E/K is a normal subgroup of Gal(E/K) if and only if L/K is a normal extension. αL (u) = α(u). so the above argument shows that L/K is normal. Then αL is the identity function on L if and only if α ∈ Gal(E/L). This shows that φ−1 θφ ∈ Gal(E/L). Thus we obtain an injective homomorphism Gal(E/K)/ Gal(E/L) −→ Gal(L/K) for which [E : K] = [L : K] =  Gal(L/K). then for every φ ∈ Gal(E/K) and v ∈ L. γ(u) ∈ E satisﬁes αγ(u) = γ(γ −1 αγ(u)) = γ(u). we have βαβ −1 ∈ Gal(E/L). αL : L −→ L.
Proof. let S(E/K) = the set of all subgroups of Gal(E/K).e. Conversely.. θ(φ(v)) = φ(v) and therefore φ−1 θφ(v) = v. →
∼ =
α Gal(E/L) −→ αL . if L/K is normal.17. For a ﬁnite Galois extension E/K. We will use the following notation. By the Monomorphism Extension Theorem 3. then αL = L since L/K is normal. Proposition.49. so for every θ ∈ Gal(E/L). Let E/K be a ﬁnite Galois extension and L/K
E/K. α −→ αL
is a group homomorphism whose kernel is Gal(E/L). hence γ(u) ∈ E Gal(E/L) = L. which shows that Gal(E/L) ▹ Gal(E/K). since γ −1 αγ ∈ Gal(E/L). Hence we can restrict α to an automorphism of L.5.4. It is easy to see that the function Gal(E/K) −→ Gal(L/K). [E : L] Hence this homomorphism is an isomorphism.
if Γ1 . then E Γ2 E Γ1 since if w ∈ E Γ2 then it is ﬁxed by every element of Γ1 (as Γ1 is a subset of Γ2 ). we indicate the subextensions in a diagram with a line going upwards indicating an inclusion. First take the 3 roots of the polynomial X 3 − 2 for √ √ √ 2 which E is the splitting ﬁeld over Q. Then the functions ΦE/K and ΘE/K are mutually inverse bijections which are orderreversing. Let L1 /K.19.18 which is closely related to Proposition 3.
. There is an immediate consequence of the Main Theorem 4. Hence ΘE/K reverses order. Since every subgroup of a ﬁnite group is a ﬁnite subset of a ﬁnite set. each of which permutes these 3 roots in the following ways given by cycle notation: α0 = (2 3). Example. α1 = (1 2 3). 4. Then Gal(E/L2 ) Gal(E/L1 ) since L1 ⊆ L2 and so if α ∈ Gal(E/L2 ) then α ﬁxes every element of L1 . Also. ΘE/K (Γ) = E Γ .Each of these sets is ordered by inclusion. When dealing with a ﬁnite Galois extension E/K.38. Figure 4. Corollary. Deﬁne two functions by ΦE/K : F(E/K) −→ S(E/K). L2 /K ∈ F(E/K) satisfy L1 /K L2 /K. Proof. We can also do this with the subgroups of the Galois group Gal(E/K) with labels indicating the index of the subgroups. This shows that ΦE/K and ΘE/K are mutually inverse and so are inverse bijections. ΘE/K (ΦE/K (L)) = ΘE/K (Gal(E/L)) = E Gal(E/L) = L. Proof. α1 . Theorem (Main Theorem of Galois Theory). Hence ΦE/K (L2 ) ΦE/K (L1 ) and so ΦE/K reverses order. 3 2 ζ3 . S(E/K) is also a ﬁnite set. ζ3 )/Q) ∼ S3 . 63
α2 = (1 3 2). α0 . Let E/K be a ﬁnite Galois extension.1 shows the Galois Correspondence for the extension of Example 3. α2 . √ As noted at the end of Example 3. Let E/K be a ﬁnite Galois extension. 4. α2 extend to automorphisms of E.
′ α2 = (1 3). In eﬀect. the Galois Correspondence inverts these diagrams. ΘE/K : S(E/K) −→ F(E/K). We know from Proposition 4.20. Γ2 ∈ S(E/K) and Γ1 Γ2 . by Proposition 4.
′ α1 = (1 2). ΦE/K (L) = Gal(E/L).14 for H ∈ S(E/K) we have ΦE/K (ΘE/K (Γ)) = ΦE/K (E Γ ) = Gal(E/E Γ ) = Γ. It = is useful to make this isomorphism explicit.16 that for an extension L/K in F(E/K).
ΦE/K
F(E/K) o
ΘE/K
/
S(E/K)
Under this correspondence. normal subextensions of E/K correspond to normal subgroups of Gal(E/K) and vice versa. α1 . 3 2 ζ3 which we number in the order ′ ′ they are listed. so is F(E/K). Since the set S(E/K) is ﬁnite. the Galois group here is Gal(Q( 3 2. Then there are only ﬁnitely many subextensions L/K E/K. Then the monomorphisms id.30. Similarly.18.20. these are 3 2. 4.
4. we have Q ER E.I ii ee eeeeee kkkk ii eeeeee kkkkk ii ee 2 ii eeeeee 2 kkk 2 eee kkk ii eeeeee √ eee √ √ 2 3 ii Q( 3 2) Q( 3 2 ζ3 ) Q( 3 2 ζ3 ) ii ii N N qq I qq ii qq ii qq ii qq 3 qq3 qq 3 Q(ζ3 ) qq Q lll qq 2lllll qq qq lll qq lll lll l Q S
ΦE/K
√ E = Q( 3 2. (1 2 3). ={id. (i) ( )E/Q agrees with the identity function if and only if ER = E.21. Galois extensions inside the complex numbers and complex conjugation When working with Galois extensions contained in the complex numbers it is often useful to make use of complex conjugation as an element of a Galois group. 4. α2 } ∼ Gal(E/Q( 2 ζ3 )) = {id. (1 3)}.
R R 64
. Of course Q(ζ3 ) is the splitting ﬁeld of X 3 − 1 over Q. ={id. (2 3)}. ( )E/Q : E −→ E. ζ3 )
k 2 kkk ww kkk wwww k kkk w
kkk 3 www kk 3 k ww kkk Gal(E/Q(ζ3 )) 3 ww kkk y ww kkk ww kkk yy w kk yy ww kkk yy √ kkk √ww √ yy 3 3 3 2 )) 3 yy Gal(E/Q( 2)) Gal(E/Q( 2 ζ3 )) Gal(E/Q( 2 ζ3 yy y yy yy 2 yy 2 2 yy
Gal(E/Q)
{id}
√ Figure 4. Proposition. Let E/Q be a ﬁnite Galois extension with E/Q C/Q. ζ3 )/Q We ﬁnd that √ 3 Gal(E/Q(ζ3 )) = {id. (1 3 2)} ▹ S3 and so Q(ζ3 )/Q is a normal extension.
Notice that {id. ( )E/Q } ∼ Z/2. Complex conjugation ( ) : C −→ C restricts to an automorphism of E over Q. α2 } ∼ Gal(E/Q( 2)) = {id. E = E ⟨ and [E : E ] = 2. Setting ER = R ∩ E. α1 .1.6. Furthermore. √ √ 2 3 3 ′ ′ Gal(E/Q( 2 ζ3 )) = {id. α1 } ∼ ={id. (1 2)}. = ( )E/Q ⟩ hence. (1 3 2)}. (1 2 3). α0 } ∼ ={id. The Galois Correspondence for E = Q( 3 2. (ii) If ER ̸= E. then ⟨ ⟩ ( )E/Q = {id.
√ Q(ζ8 )R = Q( 2). j−i
Note that this is sometimes used as the deﬁnition of sgn σ. Let u ∈ E. sgn σ = ±1. Therefore ( ) maps E into itself. and E ⟨(
)E/Q ⟩
= {u ∈ E : u = u} = ER . exactly one of the inequalities σ(i) < σ(j) or σ(j) < σ(i) must hold and the ratio (σ(j) − σ(i))/(j − i) is either positive or negative. Consider the cyclotomic extension Q(ζ8 )/Q where 1 1 ζ8 = eπi/4 = √ + √ i. the Galois group of its splitting ﬁeld E over K can naturally be thought of as a subgroup of the symmetric group Sn .
4.7.u (X) ∈ Q[X] splits over E.22. But ( ) permutes the roots of this minimal polynomial. so all of its complex roots lie in E. 2 2 From Example 4. Example. As E/Q is normal. and we easily see that [Q(ζ8 ) : Q] = 4.
65
. (i) For z ∈ C. It is easily veriﬁed that the righthand side of the following equation must have value ±1 and so (4. It is reasonable to ask when Gal(E/K) An rather than just Gal(E/K) Sn . minpolyQ. z = z if and only if z ∈ R. For each pair i. Galois groups of even and odd permutations We have seen that for a monic separable polynomial f (X) ∈ K[X] of degree n. 4.
2
C R
∞ ∞
E nnn nnn nn nnn nnn 2 n ER ee ee ee ee Q
We will usually write ( ) rather than ( )E/Q when no confusion seems likely to result.Proof. where we view the latter as permuting the roots of f (X).9 we know that √ Q(ζ8 ) = Q( 2. ⟨ ⟩ (ii) Here  ( )E/Q  = 2. i). j with 1 i<j n.3) sgn σ = ∏
1 i<j n
σ(j) − σ(i) . We ﬁrst recall an interpretation of the sign of a permutation σ ∈ Sn .
Suppose that f (X) factorizes over E as f (X) = (X − u1 ) · · · (X − un ) =
n ∏ i=1
(X − ui ). (4. .. . .
their resultant is the (m + n) × (m + n) determinant (with n rows of ai ’s and m rows of bi ’s) a0 0 . ... . . 0 0 b0 b1 . . 0 0 . . . a1 . . . .23. . . . . .. . . . . . 0 . . q(X)) = a1 a0 . . . ... Discr(X 3 + pX + q) = (−1)3 Res(X 3 + pX + q.. . . (4. bn 0 . . .. . . . The discriminant of f (X) is ∏ Discr(f (X)) = (uj − ui )2 ∈ E.. . . . . .. . . . . . . . . ... q(X) = b0 + b1 X + · · · + bn X n . 1 Discr(a0 + a1 X + a2 X 2 + X 3 ) = −27a2 + 18a0 a1 a2 + a2 a2 − 4a3 a0 − 4a3 . 0 a 0 a 1 . . . . . . . . ... . . . . . . . .24. .. .. .. . am 0 . . .. . . the ui are distinct. .. . .. .. un ∈ E are the roots of f (X). n = 2: n = 3: n = 4: Discr(a0 + a1 X + X 2 ) = −4a0 + a2 . as we have assumed that f (X) is separable. 4. 3 2 3 0 2 3 1 2 3
66
. .. . . . . . . . . . . .. . .
Here u1 . . . . . . .. .. .. .. . . . . .. . . . . . . am 0 . bn
Then if f (X) is monic with d = deg f (X). . . .. .4) Res(p(X). . . . am .. . . .. . . . .. .. . . . . . . .. . . .. ... . . . . For polynomials p(X) = a0 + a1 X + · · · + am X m . .
1 i<j n
Notice that Discr(f (X)) ̸= 0 since ui ̸= uj if i ̸= j.. 0 1 2 2 1 Discr(a0 + a1 X + a2 X 2 + a3 X 3 + X 4 ) = 18a3 a3 a2 − 6a2 a2 a0 − 192a3 a1 a2 − 27a4 1 3 1 0 1 + 144a2 a2 a2 + 144a0 a2 a2 + 256a3 − 4a3 a3 − 128a2 a2 + 16a4 a0 − 4a3 a2 3 0 1 0 3 1 2 0 2 2 1 + 18a3 a1 a2 a0 − 80a3 a1 a2 a0 − 27a4 a2 + a2 a2 a2 − 4a3 a2 a0 . . . .. . . . .. . . f ′ (X)). 0 . . 0 b0 b1 . . . .. Definition. .. .. . .. . b0 b1 . .. . . .. . . .. .
Here are some low degree examples of discriminants obtained with the aid of Maple. . . . . . ... . . . . . Remark.. . . . . .. . . . .. 0 3 Discr(f (X)) = (−1)d(d−1)/2 Res(f (X). .. . . . . ... .. . . . . . . There is an explicit formula for computing Discr(f (X)) is terms of its coeﬃcients. . . . . . bn 0 . . . . . .. 4.. . . . .. ..5) So for example. . . . .
0 . .. 3X 2 + p) q 0 = (−1) p 0 0 p q 0 p 0 0 p 3 0 p 1 0 0 3 0 0 1 0 = −4p3 − 27q 2 .. . . . .. . . .. .
If δ(f (X)) ∈ K.3) ∏ (4.6) (uσ(j) − uσ(i) ) = sgn σ
1 i<j n
∏
1 i<j n
(uj − ui ) = (±1)
∏
1 i<j n
(uj − ui ). and by Equation (4. Now let δ(f (X)) = δ(f (X))2 ∏ (uj − ui ) ∈ E. By Equation (4. 1 0
(uσ(j) − uσ(i) )2 =
∏
1 i<j n
2 (uσ(j) − uσ(i) ) . For σ ∈ Gal(E/K) σ(Discr(f (X))) = Sn . we have Discr(f (X)) ∈ K. Proposition. we have ∏
1 i<j n
Discr(X 5 + a1 X + a0 ) = 256a5 + 3125a4 . 0 4 4. On the other hand. j with i < j.
Now for each pair i. σ(uj − ui ) = uσ(j) − uσ(i) . Proof. Now consider the eﬀect of σ ∈ Gal(E/K) on δ(f (X)) ∈ E.
+ 2000a2 a3 a2 − 900a1 a3 a2 − 2500a3 a4 a1 − 50a2 a2 a2 − 900a4 a3 a2 − 27a4 a4 − 3750a3 a2 a3 0 1 3 0 0 0 4 1 2 0 4 1 0
So for example.
Hence σ(Discr(f (X))) = Discr(f (X)).
1 i<j n
Then = Discr(f (X)). so the square roots of Discr(f (X)) are ±δ(f (X)). Hence Discr(f (X)) ∈ E Gal(E/K) = K. Since E Gal(E/K) = K. if δ(f (X)) ∈ K then / K(δ(f (X))) = E Gal(E/K)∩An . Of course  Gal(E/K)/ Gal(E/K) ∩ An  = 2. this means that sgn σ = 1.
67
. σ(Discr(f (X))) = Discr(f (X)).25. For every σ ∈ Gal(E/K). σ(δ(f (X))) = sgn σ δ(f (X)) = ±δ(f (X)). Discr(X 5 + a4 X 4 + a0 ) = a3 (3125a0 + 256a5 ).n = 5:
Discr(a0 + a1 X + a2 X 2 + a3 X 3 + a4 X 4 + X 5 ) = 2250a4 a2 a3 − 36a0 a3 a3 − 128a2 a4 3 0 4 1 3 1 + 356a3 2 a2 2 a4 a1 a0 + 560a3 a2 a2 a2 − 2050a3 a2 a2 a4 a1 − 80a2 a2 a4 a1 3 − 630a3 3 a2 a4 a0 2 2 4 0 0 3 + 825a2 a2 a2 + 16a3 a3 a0 + 2000a2 a2 a3 − 6a2 a2 a3 − 128a2 a4 a2 + 16a4 a3 a0 − 4a3 a3 a2 3 2 0 3 2 4 0 2 4 1 2 4 0 2 4 2 4 1 + 108a5 a2 + 108a5 a0 − 746a3 a2 a0 a4 2 a1 2 − 27a4 a2 + 256a5 a3 − 4a3 a2 a2 + 144a3 a2 a3 3 0 2 2 1 4 0 3 2 1 2 1 + 144a2 a4 a3 + 3125a4 + 256a5 − 72a4 a2 a1 a0 + 18a3 a2 a3 a3 + 560a2 a2 a0 a2 + 16a4 a3 4 1 0 1 3 4 1 3 1 3 1 + 18a3 a2 3 a4 a1 2 − 72a3 a2 4 a4 a0 + 144a3 2 a2 a4 3 a0 2 − 192a4 4 a1 a3 a0 2 − 630a3 a2 3 a1 a0 + 24a2 3 a4 2 a1 a0 + a3 2 a2 2 a4 2 a1 2 − 6a4 3 a1 2 a3 2 a0 − 80a3 a2 2 a4 3 a1 a0 − 4a3 2 a2 3 a4 2 a0 + 2250a1 a2 a2 − 1600a3 a3 a3 − 192a4 a4 a2 − 1600a0 a3 a2 − 4a3 a3 a2 − 27a4 a2 a2 2 0 4 0 1 1 3 1 4 3 4 0 + 1020a4 2 a3 2 a0 2 a1 + 18a3 3 a2 a4 2 a0 a1 + 160a2 a4 3 a0 2 a1 + 144a2 a4 4 a0 a1 2 + 24a4 a1 2 a3 3 a0 + 1020a0 a4 a2 2 a1 2 + 160a0 a4 a1 3 a3 .6).
Sn is contained in An if and only if
4. The following result describes the Galois groups that occur and the proof introduces some useful computational techniques. we also have 4  [E : Q]. Gal(Q(δ)/Q) = Gal(E/Q)/N. where (−a + δ) (−a − δ) . since E is obtained by at most 3 successive quadratic extensions we also have [E : Q]  8. Theorem (Kaplansky’s Theorem). Example. √ √ δ(X 5 − 35X 4 + 7) = ±53 · 3 · 72 · 3 · 29 · 157 i = ±18375 13659 i ∈ Q. δ(X 5 + 20X + 16) = ±28 53 ∈ Q. Setting c = uv ∈ Q. Notice that g(X) must be irreducible since otherwise f (X) would factorize. Notice that for the Galois extension Q(δ)/Q there must be a normal subgroup N ▹ Gal(E/Q) with Q(δ) = E N . Since deg f (X) = 4. The Galois group Gal(E/K) Discr(f (X)) is a square in K. This shows that E = Q(u) and we have the following Galois tower. Let f (X) = X 4 + aX 2 + b ∈ Q[X] be irreducible. 4. v2 = . we ﬁnd that the roots of g(X) are (−a±δ)/2 ∈ Q. (i) If b is a square in Q then Gal(Q(f (X))/Q) ∼ Z/2 × Z/2.40 and 6. Let g(X) = X 2 + aX + b ∈ Q[X]. u2 = E = Q(u)
2
Q(δ)
2
Q In particular [E : Q] = 4 =  Gal(E/Q). ±v. (i) We have a2 − d 4b (uv)2 = u2 v 2 = = = b.27. 4. Proof.4. / / Then the roots of f (X) are ±u. = (ii) If b(a2 − 4b) is a square in Q then Gal(Q(f (X))/Q) ∼ Z/4.42 we obtain Discr(X 5 − 35X 4 + 7) = −4611833296875 = −33 · 56 · 74 · 29 · 157. we ﬁnd that v = c/u ∈ Q(u). / Discr(X 5 + 20X + 16) = 1024000000 = 216 · 56 .26. hence (a2 − 4b) is not a square in Q. v) which contains the quadratic extension Q(δ)/Q. = ∼ (iii) If neither b nor b(a2 − 4b) is a square in Q then Gal(Q(f (X))/Q) = D8 .
68
. 4 4 hence uv is a square root of b which is in Q. 2 2 so the splitting ﬁeld of f (X) over Q is E = Q(u. For the polynomials of Examples 6. Proposition. Setting d = (a2 − 4b) ∈ Q and taking δ to be a square root of d (so δ ∈ Q).28.8. Kaplansky’s Theorem In this section we give a detailed account of the Galois theory of irreducible rational polynomials f (X) = X 4 + aX 2 + b ∈ Q[X]. In fact.
69
. We cannot have q = 0 since this would imply that b was a square in Q. = (iii) Suppose that d. = (ii) If bd is a square in Q. Notice that c c σ(v) = =− = −u.
Q(δ)
2
Q with [E : Q] = 4 =  Gal(E/Q). Notice that if σ(v) = −u then easy calculation shows that τ σ(v) = στ (v) = u. τ. Suppose that uv ∈ Q(δ). By an easy calculation we ﬁnd that (uv)2 = b. which is a square in Q. στ } ∼ Z/2 × Z/2 = the Klein 4group. and so pq = 0. Thus we have Q(uv) ∩ Q(δ) = Q. v) = Q(u) and again we have a Galois tower E = Q(u)
2
στ = τ σ. σ(uδ) vδ so σ 2 (u) = −u. so uv ∈ E is a square root of b in E. This shows that E = Q(u.Hence there is an element σ ∈ Gal(E/Q) for which σ(δ) = −δ. σ. so we can write uvδ = c ∈ Q or equivalently v = c/(uδ) ∈ Q(u) since Q(δ) Q(u). By squaring we obtain b = (p2 + q 2 d) + 2pqδ. Since Q(δ)/Q is Galois there is an element σ ∈ Gal(E/Q) with σ(δ) = −δ and this has the eﬀect σ(u) = ±v. then uv = p+qδ for some p. A similar discussion shows that Q(uvδ) ∩ Q(δ) = Q = Q(uvδ) ∩ Q(uv). Given u we might as well choose v so that σ(u) = v. σ 2 σ 3 } ∼ Z/4 = a cyclic group of order 4. b and bd are not squares in Q. This element must also have the eﬀects σ(u) = ±v and σ(v) = ±u. We now have c c σ(u) = . This shows that Gal(Q(f (X))/Q) = Gal(E/Q) = {id. u u These satisfy σ 2 = τ 2 = (στ )2 = id = the identity. τ σ(δ) = στ (δ) = −δ. if p = 0 then b = q 2 d and so bd = (qd)2 . τ σ(u) = στ (u) = − .
hence we might as assume that σ(v) = u since if necessary we can replace our original choice by τ σ. σ. given u we might as well choose v so that σ(u) = v. then (uvδ)2 = u2 v 2 d = bd. This shows that Gal(Q(f (X))/Q) = Gal(E/Q) = {id. implying that bd was a square in Q. τ (u) = −u. There is also an element τ ∈ N for which τ (u) = −u and we also have τ (v) = −v. q ∈ Q.
uv)) Gal(E/Q) so that γ(u) = −u. Choose β ∈ Gal(E/Q(δ)) Gal(E/Q) with β(uv) = −uv. σ. σ −1 . Notice that β 2 = id. Notice that γ 2 = id. E = Q(u. αγ. Then we must have γ(v) = −v since γ(uv) = uv. Setting σ = αβ we ﬁnd σ(u) = −v and σ(v) = u. ασα = βσβ = σ −1 . δ)
www www www w
vvv vvv vvv vvv 2
Q(uv)
Q
ppp ppp 2 ppp 2 ppp
Q(uvδ)
Choose α ∈ Gal(E/Q(uv)) Gal(E/Q) so that α(δ) = −δ. α. we may assume that v = α(u) and so u = α(v).So we have a Galois tower which includes the following subﬁelds. v)
Q(δ)
r rrr rrr rrr
Q(uv. ασ −1 form a group isomorphic to the dihedral group of order 8. = and [E : Q] = 8. v)
2
Q(δ)
r 2 rrr rr r rrr
Q(uv. γ. Therefore we have Gal(Q(f (X))/Q) = Gal(E/Q) ∼ D8 . D8 . The corresponding Galois tower is E = Q(u. Then σ 2 = γ and σ has order 4. We must have either β(u) = −u or β(v) = −v. Also. so by interchanging ±δ if necessary we can assume that β(u) = −u and β(v) = v. The eight elements id. Notice that α2 = id. ασ. Choose γ ∈ Gal(E/Q(δ. By renaming −v to v if necessary. δ)
2
www www 2 www w
vvv vvv vvv vvv 2
Q(uv)
Q
p ppp 2 ppp pp 2 ppp
Q(uvδ)
70
.
(b) Let D8 be the dihedral group with the eight elements ι. and hence ∆ ∈ K. Show that ∆ = −4b3 − 27a2 . = Gal(Q(X 4 + 4X 2 + 2)/Q) ∼ Z/4. under what conditions is this always true? (b) If g(X) = X 3 + aX + b ∈ K[x] is irreducible and E = K(g(X)) is its splitting ﬁeld over K. prove that the splitting ﬁeld of f (X) over K is a ﬁnite Galois extension of K. deduce that a. b ∈ K. 3 Gal(E/K) ∼ = S3 if δ ∈ K. =
Gal(Q(X 4 + 2X 2 + 2)/Q) ∼ D8 . α.
71
. (c) Continuing with the notation and assumptions of (b). 4.4. w3 are the distinct roots of g(X) in E and let ∆ = (w1 − w2 )2 (w2 − w3 )2 (w1 − w3 )2 ∈ E.
Find all the normal subgroups of D8 . βα. and show that its discriminant is a square in Q. (a) Show that every nonabelian ﬁnite group has order at least 6. This is a revision exercise on ﬁnite groups of small order. =
Exercises on Chapter 4 4. If δ = (w1 − w2 )(w3 − w3 )(w1 − w3 ). β. β 2 = ι. βα3 satisfying α4 = ι. If f (X) ∈ K[X] is a separable polynomial. Let K be a ﬁeld for which char K ̸= 2. / [Hint: Consider K(δ) in Gal(E/K) S3 .1. βαβ = α−1 = α3 . We have the following Galois groups: Gal(Q(X 4 + 1)/Q) ∼ Z/2 × Z/2. Show that f (X) = X 3 −3X +1 ∈ Q[X] is irreducible over Q.2. 4. βα2 . explain why Gal(E/K) is isomorphic to one of the groups S3 or A3 . 3 and suppose that f (X) ∈ K[x] is a cubic polynomial. If f (X) is monic. (a) Show that there u.] E and the eﬀect on the element δ of even and odd permutations
4. v ∈ K with u ̸= 0 such that f (uX + v) = X 3 + aX + b for some a.29. b ∈ K.4. suppose that w1 . α3 . Example. Prove that the Galois group of f (X) over Q is cyclic.3. show that A if δ ∈ K. α2 . w2 .
(b) Suppose that K contains p distinct pth roots of 1. 4. √ √ √ √ Q( 23 i)(X 3 − X − 1)/Q( 23 i).28 to ﬁnd the Galois group of the splitting ﬁeld E of the polynomial X 4 + 3 ∈ Q[X] over Q.
72
. Q(i). K(X 3 − X − 1)/K for K = Q. Suppose that 0 ̸= a ∈ K and f (X) = X p − a ∈ K[X]. 4. Let K be a ﬁeld of characteristic char K = p where p > 0 is a prime. Find the Galois groups for each of the following extensions: √ √ √ √ Q(X 3 − 10)/Q. 4. Use Kaplansky’s Theorem 4. Let p > 0 be a prime. (c) Suppose that the only pth root of 1 in K is 1. Let K be a ﬁeld with char K ̸= p. Show that either f (X) is irreducible over K or it factors into p distinct linear factors over K. Q( 3 i)(X 3 − 10)/Q( 3 i). Q( 2)(X 3 − 10)/Q( 2). describe the remaining roots and show that L contains p distinct pth roots of 1.6. Q( 5).5. Q( 5 i).7. (a) Show that f (X) has p distinct roots in L. If u ∈ L is one such root.4. Show that either f (X) is irreducible over K or it has a root in K. Determine all the subextensions F E for which F/Q is Galois. Show that if f (X) has no root in K then it is irreducible over K.8. Let L/K where L is a splitting ﬁeld for f (X) over K. Suppose that 0 ̸= a ∈ K and f (X) = X p − a ∈ K[X].
. Then E is ﬁnite if and only if E/F is ﬁnite and then E = F [E:F ] . Corollary. Lemma. 5.35 Θpd (X) must have exactly pd distinct roots in Fp . Fpd is a ﬁnite subﬁeld of Fp with pd elements and F0d = F× . Let Fpd = {u ∈ Fp : Θpd (u) = 0} ⊆ Fp .3. p
= 1. Our ﬁrst goal is to count the elements of K. Our next task is to show that for each power pd there is a ﬁnite ﬁeld with pd elements. . hence by Proposition 3. Corollary. upd −1 . Here is a more general result. u1 . . A thorough account of ﬁnite ﬁelds and their applications can be found in [6]. If d = dimF V < ∞. say 0. Throughout this chapter we will assume that K is a ﬁeld of prime characteristic p = char K > 0. Finite ﬁelds If K is a ﬁnite ﬁeld.4. . Let K be a ﬁnite ﬁeld. . and each root is separable over Fp . 5. . Let F be a ﬁnite ﬁeld and E/F an extension. For each d 1. Proof. Clearly there are exactly q d such expressions. Conversely. if V is ﬁnite then any basis has ﬁnitely many elements and so dimF V < ∞. . td ∈ F . vd we can express each element v ∈ V uniquely in the form v = t1 v1 + · · · + td vd . We start with the algebraic closure Fp of Fp and consider the polynomial Θpd (X) = X p − X ∈ Fp [X].55 every root of Θpd (X) in Fp is simp ple. Then dimF V < ∞ if and only if V is ﬁnite in which case V  = q dimF V . Proposition. . . . 5. . Let F be a ﬁnite ﬁeld with q elements and let V be an F vector space. Then K/Fp is ﬁnite and K = p[K:Fp ] . Furthermore. p pd
73
.1. Then in Fp [X] we have X p − X = X(X − u1 ) · · · (X − upd −1 ). Notice that Θ′ d (X) = −1.
5. especially ﬁnite ﬁelds. containing the prime subﬁeld Fp . the extension Fpd /Fp is a separable splitting ﬁeld. Notice that u ∈ F0d if and only if up p
d −1 d d
F0d = {u ∈ Fpd : u ̸= 0}.CHAPTER 5
Galois extensions for ﬁelds of positive characteristic
In this chapter we will investigate extensions of ﬁelds of positive characteristic. then for a basis v1 . where t1 . so V  = q d . 5. Therefore by Corollary 1. . then K is an Fp vector space.2.1.
(uv)p − uv = up v p − uv = uv − uv = 0. d d each element u ∈ F × has order dividing pd − 1.6.7. the image of the resulting monomorphism must be Fpd . subﬁelds both have p (iii) Apply the Monomorphism Extension Theorem 3.
d d
(i) Fpd Fp is the splitting subﬁeld for each of the polynomials X p − X and X p −1 − 1 over Fp . so for a. hence F ∼ Fpd . Let d 1.
74
.10. By inspection. hence F0d = F× . so Fpd /Fp is a ﬁnite extension. Consider the polynomial X 4 − X ∈ F2 [X]. d ∈ F2 we have (a + bw)(c + dw) = ac + (ad + bc)w + bdw2 = (ac + bd) + (ad + bc + bd)w. Fp have pd elements. The ﬁnite subﬁeld Fpd
Fp is called the Galois ﬁeld of order pd . By (ii). Proposition. 5. Of course. To calculate products we use the fact that w2 = w + 1. so by Lagrange’s Theorem. Let K be a ﬁnite ﬁeld of characteristic p. Corollary. Fp1 = GF(p1 ) = GF(p) = Fp and [Fpd : Fp ] = d.Proof. 5. Furthermore. In any ﬁeld the nonzero elements are always invertible.8. therefore up −1 = 1 and so up = u. If u. Now X 2 + X + 1 has no root in F2 so it must be irreducible in F2 [X]. Definition. = Proof. equality follows since these d elements. in the ring F2 [X] we ﬁnd that X 4 − X = X 4 + X = X(X 3 + 1) = X(X + 1)(X 2 + X + 1). d The nonzero elements of Fpd are the roots of X p −1 − 1. the other being w + 1 since the sum of the roots is the coeﬃcient of X. This group is abelian and has pd − 1 elements. Hence Fpd Fp . b. Its splitting ﬁeld is a quadratic extension F2 (w)/F2 where w is one of the roots of X 2 + X + 1. This tells us that every element of F4 = F2 (w) can be uniquely expressed in the form a + bw with a. Notice that the nonzero elements of F form a group F × (ii) Let F under multiplication. But this means every element of F is a root of Θpd (X) and so F Fpd . b ∈ F2 . (ii) Fpd Fp is the unique subﬁeld with pd elements. if u ̸= 0 then up −1 = 1 and so u has multiplicative inverse up −2 . = It is worth noting the following consequence of this result and the construction of Fpd . 5. Example. Then K/Fp is a ﬁnite Galois extension. so Fpd is also the splitting subﬁeld for this polynomial.49 for K = L = Fp and M = F . it is the splitting subﬁeld. Notice that Fp Fpd .5. (iii) If F is any ﬁeld with pd elements then there is a monomorphism F −→ Fp with image Fpd . (i) As Fpd consists of exactly the roots of Θpd (X) in Fp .
The notation GF(pd ) is often used in place of Fpd . p pd
d d d d d d d d d d
5. c. (u + v)p − (u + v) = (up + v p ) − (u + v) = (up − u) + (v p − v) = 0. therefore F ∼ Fpd . v ∈ Fpd then by the Idiot’s Binomial Theorem 1.
Proof. so if u ∈ F3 is a root of X 2 + 1 then F3 (u)/F3 has degree 2 and F3 (u) = F9 .10 now implies that Fp (u) = Fpn for some n. b ∈ F3 . The group of units F× in Fpd is cyclic. ∪ Fp = Fpn . Similarly. Let u ∈ Fp . the discriminant of X 2 − X − 1 is 1 − 4(−1) = 5 = 2 = u2 and its roots are (−1)(1 ± u) = −1 ± u. each element u ∈ Fp is separable over Fp . The diagram of subﬁelds for Fp24 can be seen in Figure 5. Let Fpm and Fpn be two Galois ﬁelds of characteristic p. Consider the polynomial X 9 − X ∈ F3 [X].2. By inspection.10. The discriminant of X 2 + X − 1 is 1 − 4(−1) = 5 = 2 = u2 . Then u is algebraic over Fp and the extension Fp (u)/Fp is ﬁnite. So X 2 + X − 1 and X 2 − X − 1 are also quadratic irreducibles in F3 [X]. 5. 5. We will require a useful fact about Galois ﬁelds. 5.12.9. Theorem.11. write n = km with k up = up
n mk
1. The algebraic closure of Fp is the union of all the Galois ﬁelds of characteristic p. Every element of F9 can be uniquely expressed in the form a + bu with a. then by Corollary 5. Then we have F9 = F3 (u) = F3 (1 ± u) = F3 (−1 ± u).5. Proof.
n 1
Furthermore. in the ring F3 [X] we ﬁnd that X 9 − X = X(X 8 − 1) = (X 3 − X)(X 2 + 1)(X 2 + X − 1)(X 2 − X − 1).
= up
m(k−1)
= · · · = up = u.2. Hence by Corollary 5. so its roots are (−1)(−1 ± u) = 1 ± u. Proposition.7. Multiplication is carried out using the relation u2 = −1 = 2. so
m m(k−1)
m
= (up )p Fpn . If m  n. Proposition. If Fpm Fpn . There are two issues we can now clarify. Example.1 which shows extensions with no intermediate subextensions.
m
Hence u ∈ Fpn and therefore Fpm
This means that we can think of the Galois ﬁelds Fpn as ordered by divisibility of n. We can ﬁnd their roots in F9 using the quadratic formula since in F3 we have 2−1 = (−1)−1 = −1. Fp (u) Fp is a ﬁnite subﬁeld. hence X 2 + 1 ∈ F3 [X] is irreducible. so m  n. Then for u ∈ Fpm we have up = u. pn = (pm )[Fpn :Fpm ] = pm[Fpn :Fpm ] . Proposition 5. Then Fpm Fpn if and only if m  n. Notice that X 2 + 1 has no root in F3 . The separability statement follows from Corollary 5. Let us ﬁnd an irreducible polynomial of degree 2 in F3 [X]. pd
75
.
its order in the group F× is (pd − 1). Remark.1. pd pd 5. 5. In Example 5.
76
.75 which we had previously only established for inﬁnite ﬁelds. Proof. Then every element of Fpnd can be expressed as a polynomial in w. Fpnd = Fpd (u) for some u ∈ Fpnd . Then every ﬁnite subgroup U Proof. so Fpnd Fpd (w) Fpnd .17. Since 9 φ(8) = 4. although as we will in the next result. some authors use the term primitive element for what we have called a primitive root. Use Corollary 1. hence ⟨w⟩ = F× . We record a fact that is very important in Number Theory.e. Fpnd has a primitive root w say.35 and Lemma 1. The extension of Galois ﬁelds Fpnd /Fpd is simple. Example. Proposition.46..12. there are four primitive roots and these are the roots of the polynomials X 2 + X − 1 and X 2 − X − 1 which we found to be ±1 ± u. 5. i. 5.8 we ﬁnd that F4 = F2 (w) has the two primitive roots w and w + 1.9 we have F9 = F3 (u) and F× is cyclic of order 8. Example. Indeed. 5. This implies that Fpnd = Fpd (w).15. Definition.20. the polynomial X 2 + 1 ∈ Fp [X] has two roots in Fp . By Proposition 5. The subﬁelds of Fp24 This is a special case of a more general result about arbitrary ﬁelds. Let K be a ﬁeld.. In Example 5. Proposition. every primitive root is indeed a primitive element in our sense! 5.14. but that conﬂicts with our usage.16.19. i. Unfortunately the word primitive has two confusingly similar uses in the context of ﬁnite ﬁelds. unity.13.e. (i) If p ≡ 1 (mod 4). w ∈ F× is called a primitive root if it is a primitive (pd − 1)th root of pd K × is cyclic. This completes the proof of the Primitive Element Theorem 3.18. Proposition. 5.Fp8
{ {{ {{ { {{ { {{ {{ { {{ {{ {{ {{ {{
Fp24
Fp12
Fp4
Fp6
gg gg gg gg { {{ {{ { {{
Fp2
gg gg gg gg
Fp3
Fp
Figure 5. Let p > 0 be an odd prime. Remark.
5.
23. Galois groups of ﬁnite ﬁelds and Frobenius mappings Consider an extension of Galois ﬁelds Fpnd /Fpd .
d
d
d
so Fd is a ring homomorphism. so Gal(Fpnd /Fpd ) = ⟨Fd ⟩. Fpd contains a primitive nth root of unity if and only if pd ≡ 1 (mod n) and p n. The order of Fd is n. hence
1. Then Fd is invertible with inverse F−1 = Fd . For d Fd : Fp −→ Fp . Proposition. By Proposition 5. (i) We have 4  (p − 1) = F× .
then every element u ∈ Fpnd satisﬁes the equation Fk (u) = u which expands to up d u ∈ Fpkd . which is p p impossible since p − 1 ≡ 2 (mod 4). Suppose the order is k with k n. 5. so Fd ∈ Gal(Fpnd /Fpd ).73.
77
. hence Fn = id. Here is a generalization of Proposition 5. 5.20.6(i).(ii) If p ≡ 3 (mod 4) the polynomial X 2 +1 ∈ Fp [X] is irreducible. Proposition. Frobenius mappings exist on the algebraic closure Fp . Let d 1. But this can only be true if k = n. = Proof. (ii) If v ∈ Fp is a root of X 2 + 1 then v has order 4 in F× . so Fd ﬁxes the elements of Fpd . The (relative) Frobenius mapping for the extension Fpnd /Fpd is the funcd tion Fd : Fpnd −→ Fpnd given by Fd (t) = tp . this extension is Galois and  Gal(Fpnd /Fpd ) = [Fpnd : Fpd ] = n. Proposition.
d d d × ×
Fd (uv) = (uv)p = up v p . notice that the composition power Fn = Fd ◦ · · · ◦ Fd (with n factors) satisﬁes d Fn (t) = tp d
nd d
=t
n−1 for all t ∈ Fpnd .21. Also. But then 4  (p − 1) = F× . v ∈ Fpnd . 5. the p p order of uFp /4 is 4. To see that Fd is an automorphism. so Fp2 ∼ Fp [X]/(X 2 +1). Fd (u) = u if and only if u ∈ Fpd . Proof. we have the identities Fd (u + v) = (u + v)p = up + v p . 5.22.7 and Proposition 3. 5. The relative Frobenius mapping Fd : Fpnd −→ Fpnd is an automorphism of Fpnd that ﬁxes the elements of Fpd . For u. the cyclic group generated by Fd . so if u ∈ F× is a generator of this cyclic group. hence this is a root of X 2 + 1 (the other root is −uFp /4 ).
d
kd
= u. consider the function
(i) Fd : Fp −→ Fp is an automorphism of Fp which ﬁxes the elements of Fpd . Fd (t) = tp . Definition. This also shows d d that the order of Fd in the group AutFpd (Fpnd ) is at most n. for u ∈ Fpd we have Fd (u) = up = u.24. We next introduce an important element of the Galois group Gal(Fpnd /Fpd ). Corollary 5. In fact for u ∈ Fp .2.
1. then Fk = Fkd .2. For u ∈ Fpnd we have Fk (u) = up .2 shows the subgroup diagram corresponding to the lattice of subﬁelds of Fp24 shown in Figure 5. The subgroups of the Galois groups of Fp24 /Fp
78
. For k  n.(ii) The restriction of Fd to the Galois subﬁeld Fpdn agrees with the relative Frobenius mapping Fd : Fpnd −→ Fpnd . so AutFpd (Fp ) is inﬁnite. Proof. ⟨ ⟩ In Gal(Fpnd /Fpd ) = ⟨Fd ⟩. (ii) If k 1. Proposition. hence Fk (u) = u if and only if u ∈ Fpdk . This is left as an exercise. Hence in the automorphism group AutFpd (Fp ).
dk
l lll lll ll ll lll ⟨ 3⟩ ⟨ 2⟩ l F F l l lll ll lll lll ⟨ ⟩ llll ⟨ 4⟩ F F6 ⟨ ⟩ ⟨ ⟩
Gal(Fp24 /Fp ) = ⟨F⟩ ∼ Z/24 =
F12 F8 ⟨ 24 ⟩ F = {id}
Figure 5. d d Figure 5. for each k with k  n there is the cyclic subgroup Fk of order d ⟨ ⟩  Fk  = n/k. The Frobenius mapping F = F1 is often called the absolute Frobenius mapping since it exists as an element of each of the groups AutFp (Fp ) and AutFp (Fpn ) = Gal(Fpn /Fp ) for every n 1.25. the ﬁxed subﬁeld of Fk in Fpnd is Fpnd = Fpdk . Fd has inﬁnite d order. d Fpnd
n/k
F
⟨ ⟩
Fk d pnd
= Fpdk
k
Fpd Proof. d ⟨ ⟩ ⟨Fk ⟩ d 5.
26.27. Proposition. Proposition.24(i). If we modify TFpnd /Fpd to have codomain Fpd .
5.
= up up up · · · up = uu u = N(u). TrFpnd /Fpd is surjective since its codomain has dimension 1. pnd pd NormFpnd /Fpd (u) = uup up · · · up
d 2d (n−1)d
. TFpnd /Fpd (u) ∈ Fpd . consider the function TFpnd /Fpd : Fpnd −→ Fpnd deﬁned by TFpnd /Fpd (u) = u + up + up Notice that Fd (TFpnd /Fpd (u)) = up + up = up + up
d d 2d d 2d
+ · · · + up
(n−1)d
= u + Fd (u) + F2d (u) + · · · + F(n−1)d (u). The trace and norm mappings For an extension of Galois ﬁelds Fpnd /Fpd . By redeﬁning the codomain we obtain the relative norm NormFpnd /Fpd : F× −→ F× . so Fpd linearity follows from the formula tu + (tu)p + (tu)p + · · · + (tu)p
d 2d (n−1)d d
= tu + tup + tup
d
2d
+ · · · + tup
(n−1)d
. Consider the function N : F× −→ F× pnd pnd for which N(u) = uup up · · · up Then we have Fd (N(u)) = up up up · · · up
d 2d 3d d 2d 3d nd d 2d (n−1)d
= u Fd (u) F2d (u) · · · F(n−1)d (u). Clearly TrFpnd /Fpd is additive. TrFpnd /Fpd (u) = u + up + up + · · · + up
d 2d (n−1)d
. This means that ker TrFpnd /Fpd cannot be the whole of Fpnd . The relative norm NormFpnd /Fpd is a surjective group homomorphism. N(u) ∈ Fpd .24(i).3. For t ∈ Fpd we have tp = t.
pd p2d
(n−1)d
u
u
p3d
···u
p(n−1)d
So by Proposition 5. There is a multiplicative version of this construction.
+ up + up
3d
+ · · · + up + · · · + up
nd
2d
3d
(n−1)d
+ u = TFpnd /Fpd (u). The relative trace TrFpnd /Fpd is a surjective Fpd linear mapping and whose kernel is an Fpd vector subspace of dimension n − 1. Proof. notice that TrFpnd /Fpd (u) = 0 if and only if u is a root of the polynomial X + Xp + Xp
d 2d
+ · · · + Xp
(n−1)d
∈ Fpd [X]
which has degree p(n−1)d and so has at most p(n−1)d < pnd roots in Fpnd . we obtain the relative trace TrFpnd /Fpd : Fpnd −→ Fpd .
To see that TrFpnd /Fpd is surjective.
So by Proposition 5.
79
.5.
5.
4. 1 + pd + · · · + p(n−1)d = pnd − 1  ker NormFpnd /Fpd 
d +···+p(n−1)d
− 1 ∈ Fpd [X].
Exercises on Chapter 5 5. Multiplicativity is obvious. The kernel of NormFpnd /Fpd consists of the roots in Fpnd of the polynomial X 1+p so  ker NormFpnd /Fpd  Hence  im NormFpnd /Fpd  = Since im NormFpnd /Fpd F× .3. pd − 1 pd − 1. Suppose that d d ga (X) = X p − X − a ∈ K[X]. [Hint: show that if u ∈ E is a root of ga (X) = in an extension E/K. (b) X 8 − 1 ∈ F5 [X].5. 5. the splitting ﬁelds of f (X) and g(X) over Fpd agree.13 also applies to an integral domain in place of a ﬁeld. 5. and K/Fpd is an extension.1. show that ga (X) is irreducible over K if and only if it has no root in K. let
(a) If the polynomial ga (X) is irreducible over K. Let p > 0 be a prime. 5. Find the splitting ﬁeld of f (X). 1. explain why ga (X) can never be irreducible over K.20 if we try to take p = 2. then so is u + t for every t ∈ Fp .] (b) If d = 1. For a ∈ K. Let f (X) ∈ Fpd [X] be an irreducible polynomial with deg f (X) = n.
(a) Consider {±1} = {1. pd pd − 1. Deduce that for any other irreducible polynomial g(X) ∈ Fpd [X] with deg g(X) = n.6. −1} as a group under multiplication. d 1 and write q = pd . in each case ﬁnd a primitive root of this Galois ﬁeld: (a) X 8 − 1 ∈ F41 [X]. pnd − 1 . (d) X 8 − 1 ∈ F2 [X]. show that the splitting ﬁeld E of ga (X) over K is separable and Gal(E/K) ∼ Fpd .
5.Proof. Show that there is a unique group homomorphism λq : F× −→ {±1} which is characterized by the requirement q × . (c) If K is a ﬁnite ﬁeld and d > 1. 5. Find the smallest Galois ﬁelds containing all the roots of the following polynomials. Let p be an odd prime. 5. Let w ∈ F× be a primitive root. Show that Proposition 5.2. If ℓ < d. Deduce that degFp w = d / pℓ pd and d  φ(pd − 1). Is λ always that for every u ∈ Fq q q q surjective?
80
. What happens to Theorem 5. show that w ∈ F× . (c) X 8 − 1 ∈ F11 [X]. we also have pd  im NormFpnd /Fpd  therefore im NormFpnd /Fpd = F× . λ (u) = 1 if and only if u = v 2 for some v ∈ F× .7.
(e) What can you say about the case p = 2?
81
. show that Σq ∩ (t − Σq )
1.(b) Consider the set of all squares in Fq . (c) If t ∈ Fq .
Deduce that every element of Fq is either a square or can be written as the sum of two squares. (d) Deduce that the equation x2 + y 2 + z 2 = 0 has at least one nontrivial solution in Fq . Σq = {u2 ∈ Fq : u ∈ Fq } ⊆ Fq . Deduce that if t ∈ Fq then the set t − Σq = {t − u2 ∈ Fq : u ∈ Fq } has t − Σq  = (q + 1)/2 elements. Show that the number of elements of Σq is Σq  = (q + 1)/2.
.
It is interesting to compare the proof below with others which use the topology of the plane and circle or Complex Analysis. so C/R is algebraic. so suppose not. This shows that Gal(E/R) = P . Complex Analysis and Algebraic Geometry.CHAPTER 6
A Galois Miscellany
In this chapter we will explore some miscellaneous topics in Galois Theory.
83
. The Primitive Element Theorem 3. Therefore  Gal(E/C) = 1 and E = C. The ﬁeld of complex numbers C is algebraically closed and R = C.75 allows us to write E P = R(v) for some v whose minimal polynomial over R must also have odd degree. Galois Theory has always been an important tool in Number Theory and Algebra. But from known properties of C (see Proposition 3.u (X)(X 2 + 1) over R. our proof only uses the connectivity of the real line (via the Intermediate Value Theorem) together with explicit calculations in C involving square roots. 6. so in C[X] we have p(X)  minpolyR.1. Theorem (The Fundamental Theorem of Algebra). As C/R is a Galois extension. we have 2  [E : R] and so 2   Gal(E/R). We know that [C : R] = 2.1.29). Then any root u of p(X) in the algebraic closure C is algebraic over R. The splitting ﬁeld of p(X) over C is contained in the splitting ﬁeld E of minpolyR. Since C E. But by the Intermediate Value Theorem. So we cannot have an irreducible quadratic polynomial in C[X].u (X). Let p(X) ∈ C[X] be irreducible. so irreducibility implies that v has degree 1 over R and therefore E P = R. For the ﬁxed subﬁeld of P . every quadratic aX 2 + bX + c ∈ C[X] has complex roots (because we can ﬁnd square roots of every complex number). every real polynomial of odd degree has a real root. This proof is essentially due to Gauss but he did not use the historically more recent Sylow theory. Ring Theory and such diverse areas as Differential Equations. From the theory of 2groups. Proof. A proof of the Fundamental Theorem of Algebra We will prove the Fundamental Theorem of Algebra for the complex numbers C. hence Gal(E/R) is a 2group. we can consider the normal subgroup Gal(E/C)▹Gal(E/R) for which  Gal(E/R) = 2  Gal(E/C). there is a normal subgroup N ▹ Gal(E/C) of index 2. we have [E P : R] =  Gal(E/R) . We must show that  Gal(E/C) = 1. P 
which shows that E P /R has odd degree. stimulating the development of subjects such as Group Theory. Many of the ideas introduced in this chapter are of great importance in these and other mathematical areas. so we can consider the Galois extension E N /C of degree 2. Historically. Now consider a 2Sylow subgroup P Gal(E/R) and recall that  Gal(E/R)/P  is odd. 6.
n − 1 and t gcd(t. Since the complex roots of Φn (X) are the powers ζn with t = 1. this is a cyclic subgroup of order n. To see this. ⟨ζn ⟩ Z(ζn )× .. π gives an injective r group homomorphism.. it is a ﬁnite ﬁeld. t • Gal(Q(ζn )/Q) ∼ (Z/n)× . say Fpd for some d. π ′ : ⟨ζn ⟩ −→ F× .
so ζn − 1 ∈ P or r ∈ P since maximal ideals are prime. we can consider the eﬀect of the absolute Frobenius map F : Fpd −→ Fpd
t on ζ n = ζn . Q(ζn ) is the splitting ﬁeld of Φn (X) over Q and indeed Q(ζn ) = Q(ζn ) whenever t t is a primitive nth root of unity. Inside the group of units of Z(ζn ) is the subgroup of powers of ζn . 2. Then
• Q(ζn ) = Q[X]/(Φn (X)). Let n
2.
Then Z(ζn ) is a subring of Q(ζn ) and so is an integral domain. 84 t t t
. In fact.43. where the element tn ∈ (Z/n)× acts on Q(ζn ) by tn · ζn = ζn . By elementary Group Theory we can assume that r  n and so p r. Writing u = π ′ (u).. it was claimed that the irreducible polynomial over Q which has ζn as a root was the nth cyclotomic polynomial ∏ t Φn (X) = (X − ζn ). Consider
r Z(ζn ) = {a0 + a1 ζn + · · · + ar ζn : r
0. then the quotient ring Z(ζn )/P is a ﬁeld of characteristic p. tp F(ζ n ) = (ζ n )p = ζn . Let p > 0 be a prime which does not divide n. • [Q(ζn ) : Q] = φ(n). n) = 1.ζn (X) = minpolyQ. the standard primitive nth root of 1 in C. Let P ▹ Z(ζn ) be a maximal ideal which contains p. Cyclotomic extensions We begin by discussing the situation for cyclotomic extensions over Q using material discussed in Section 1. .
we see that n ∈ P and hence p  n. The main step in the proof has the above properties and so ζn t is to show that Φn (X) ∈ Z[X] is irreducible.n)=1
6. Recalling that
n−1 ζn + · · · + ζn + 1 = 0. Theorem. hence / ζn − 1 ∈ P . =
t Proof.3. .n−1 gcd(t. On factoring we have
r−1 (ζn − 1)(ζn + · · · + ζn + 1) ≡ (ζn − 1)r
(mod P ). n − 1. .
t=1. . . We claim that when restricted to ⟨ζn ⟩. Let π : Z(ζn ) −→ Fpd be the quotient homomorphism. . But Z ∩ P = (p) and so r ∈ P . .6. thus contradicting our original assumption on n. So π ′ is injective.2. suppose that π ′ (ζn ) = 1 for some r = pd r 1. aj ∈ Z} ⊆ Q(ζn ). Its group of units contains the cyclic subgroup ⟨ζn ⟩ of order n..2. therefore
Φn (X) = minpolyQ. In Theorem 1. Let ζn = e2πi/n . To do this we will show that every power ζn as above is actually a Galois conjugate of ζn over Q. . then ζn − 1 ∈ P .ζn (X) t and hence Φn (X) is irreducible.
Hence t ζn is conjugate to ζn in the extension Q(ζn )/Q. Theorem.
p 1 f (ζn1 )
p 1 f (ζn1 )
r
̸= 0 in Z[ζn ]. this shows that minpolyQ. For n > 2. where the p are primes with 2 t = p1 · · · pm p1 < · · · < pm and rj 1 with. . . Furthermore.3. 2 2 Complex conjugation ﬁxes each of the real numbers cos(2πk/n) for k = 1.ζn (X)
t for some monic polynomial f (X) ∈ Z[X] and f (ζn ) = 0. it is elementary to show that cos(2πr/n) ̸= cos(2π/n) unless r ≡ 1 (mod n).
Gal(Q(ζn )/Q) ∼ Z/n× .
85
. . Now consider a maximal ideal P1 ▹ Z[ζn ] containing p1 . consider the cyclotomic extension Q(ζn )/Q where ζn = e2πi/n . When j = m. . . By separability and
1 r
p
r1
̸= 0 and so conjugate to ζn . . we ﬁnd that ζ n is conjugate to ζ n1 . 2
)⟩
= Q(ζn + ζ n ) = Q(cos(2π/n)). 2
1 1 −1 cos(2π/n) = (ζn + ζ n ) = (ζn + ζn ). we obtain
1 cos θ = (eθi + e−θi ). n − 1 and gcd(t. This shows that minpolyQ. Making use of the identities
eθi = cos θ + sin θ i. Reducing modulo P1 and working
the fact that the reduction map π1 : Z[ζn ] −→ Fpd1 is injective on the powers of ζn .tp This shows that in the Galois extension Fpd /Fp . Continuing in this fashion. Hence ⟨ ⟩ ( ) = {id. Since j gcd(t. . Consider the prime power factorization r1 rm . n) = 1 we also have pj n s. The residue class of r acts by sending cos(2π/n) to cos(2πr/n).ζn (ζn1
p
r r1 r2 p2 ···pj j
)=0
t and so ζn1 j is conjugate to ζn . for each j = 1.ζn (ζn1 ) = 0 and so ζn1 is
p
r1
p
r1
p
r1
Repeating this argument starting with ζn1 and using the prime p2 we ﬁnd that minpolyQ. Complex conjugation corresponds to the residue class of −1 ≡ n − 1 (mod n). . Suppose there is a factorization t
k
t
Φn (X) = f (X) minpolyQ.ζn (ζn1 and so ζn
r r p11 p22
p
r1 r2 p2
)=0
is conjugate to ζn . Q(ζn )R = Q(ζn )⟨( and [Q(cos(2π/n)) : Q] = Proof. ( )} = Gal(Q(cos(2π/n))/Q). . . . . ζ n is conjugate to ζn . Recall that φ(n) .ζn (ζn ) = 0. n − 1. Then Q(ζn )R ̸= Q(ζn ). =
r where the residue class of r acts by sending ζn to ζn . Now let t = 1. n) = 1. we ﬁnd that
1
in the resulting extension Fpd1 /Fp1 . m we have minpolyQ. by iterating this we ﬁnd tp that ζ n is conjugate to every power of the form ζn .
p
r r1 ···p j
6. 2.
3.28 applies to the polynomial minpolyQ. By Theorem 1. Gal(Q(ζ24 )/Q) ∼ Z/2 × Z/2 × Z/2. i) : Q] = 8.
86
. 4 giving 16 cos4 (π/12) − 16 cos2 (π/12) + 1 = 0.ζn (X) = X 2 − 2 cos(2π/n)X + 1. Also.cos(π/12) (X). It is easy to check that √ √ [Q( 2. hence Q( 2. ζ24 = 2 2 2 2 √ √ and all of these numbers are in Q(ζ24 ). and so [Q(cos(2π/n)) : Q] = φ(n)/2. ζ24 = − + i.
6.cos(π/12) (X). so we also have (6. √ √ √ 2 2 1 3 8 6 3 + i. i) Q(ζ24 ).1) minpolyQ(cos(2π/n)). Example. = We also have cos(2π/24) = cos(π/12) ∈ Q(ζ24 ).Thus we have Q(ζn )⟨( )⟩ = Q(cos(2π/n)). Note that case (i) of Kaplansky’s Theorem 4.43 we have [Q(ζ24 ) : Q] = 8. Then 16X 4 − 16X 2 + 1 = 16 minpolyQ. Notice that ζn is a root of the polynomial X 2 − 2 cos(2π/n)X + 1 ∈ Q(cos(2π/n))[X]. 3. Since √ 3 cos(2π/12) = cos(π/6) = . = Proof.4. which implies that √ √ Q(ζ24 ) = Q( 2. We have [Q(ζ24 ) : Q] = φ(24) = 8 and Gal(Q(ζ24 )/Q) ∼ Z/2 × Z/2 × Z/2. ζ24 = i. i). 2 √ 3 2 cos (π/12) − 1 = 2
2
Using this we ﬁnd that
we have
and so 3 4 cos4 (π/12) − 4 cos2 (π/12) + 1 = . For this example. it is an interesting exercise to ﬁnd them all together with their ﬁxed subﬁelds. 3. Gal(Q(ζ24 )/Q) has 23 − 1 = 7 subgroups of each of the orders 2 and 4.
Now
−1 −4 −2 4 2 (ζ24 + ζ24 )4 = (ζ24 + ζ24 ) + 4(ζ24 + ζ24 ) + 6. then the splitting ﬁeld of X n − 1 over K is K(ζ). we obtain
8 4 ζ24 − ζ24 + 1 = 0. T4 (X) = 8X 4 − 8X 2 + 1. The polynomial Φn (X) has integer coeﬃcients. If char K n. However.
Similarly. T5 (X) = 16X 5 − 20X 3 + 5X. 6.e.
and so 16 cos4 (π/12) − 16 cos2 (π/12) + 1 = 0.
−1 −2 2 (ζ24 + ζ24 )2 = ζ24 + ζ24 + 2.. The minimal polynomial for cos(π/12) can also be found as follows. The Galois group Gal(K(ζ)/K) is isomorphic to a subgroup of (Z/n)× . The polynomial that expresses cos nθ as a polynomial in cos θ is the nth Chebsyhev polynomial of the ﬁrst kind Tn (X) ∈ Z[X]. Now let K be a ﬁeld with characteristic char K n. Here are the ﬁrst few of these polynomials: T2 (X) = 2X 2 − 1. This method will work for any n where φ(n) is even. T6 (X) = 32X 6 − 48X 4 + 18X 2 − 1. so we can view it as an element of K[X] since either Q K or Fp K and we can reduce the coeﬃcients modulo p.6. we can still describe the splitting ﬁeld of X n − 1 over K and its Galois group. T7 (X) = 64X 7 − 112X 5 + 56X 3 − 7X.
so
−2 −1 2 ζ24 + ζ24 = (ζ24 + ζ24 )2 − 2.7. In either case it can happen that Φn (X) factors in K[X]. hence since Φ24 (X) = X 8 − X 4 + 1.
These form a system of orthogonal polynomials which can be computed in Maple using the command orthopoly[T](n.5. Theorem.X).
giving
−4 4 (ζ24 + ζ24 ) − 1 = 0. −4 Then after multiplying by ζ24 we have −4 4 ζ24 − 1 + ζ24 = 0. T3 (X) = 4X 3 − 3X.
hence
−4 −1 −2 4 2 ζ24 + ζ24 = (ζ24 + ζ24 )4 − 4(ζ24 + ζ24 ) − 6. We have Φ24 (ζ24 ) = 0. i. where ζ ∈ K is a primitive nth root of unity. Remark. 6.
87
. when n > 2.6. hence it is abelian with order dividing φ(n).
Combining these we have
−1 −1 (ζ24 + ζ24 )4 − 4(ζ24 + ζ24 )2 + 1 = 0. Remark.
Then χ1 . . The splitting ﬁeld of (X 2 + 1) over F3 is F9 = F3 (ζ). = = (ii) By Proposition 5. 6.20. .9. Given an automorphism α : K −→ K. X 4 − 1 splits over F5 . . . For example. tn ∈ K. Let χ1 .20. Example. Given any ring homomorphism φ : R −→ K we obtain a character of R× in K by restricting φ to a map χφ : R× −→ K × . where ζ 2 + 1 = 0. In this deﬁnition. we know that Φn (ζ) = 0.12. This generator clearly extends to an automorphism of F9 (T ) which ﬁxes T . (i) The splitting ﬁeld of X 4 − 1 over F3 (T ) is F9 (T ) and Gal(F9 (T )/F3 (T )) ∼ (Z/4)× ∼ Z/2. Let E/K be a Galois extension and σ ∈ Gal(E/K).
88
. Working in K. Hence Gal(K(ζ)/K) is isomorphic to a subgroup of Gal(Q(ζn )/Q) ∼ (Z/n)× which implies that it is abelian and its order divides φ(n). t1 χ 1 + · · · + tn χ n = 0 =⇒ t1 = · · · = tn = 0. .21. Example.13.8. . = 6. X 4 − 1 splits over F5 (T ) and the Galois group Gal(F5 (T )/F5 (T )) is trivial. 6. so (X 2 + 1) splits over F9 (T ).20. . . Example. Gal(F9 /F3 ) ∼ (Z/4)× ∼ Z/2. 6. 6. χn are not linearly independent then they are linearly dependent. 6. Also. if K = Fpd (T ) then the value of d is the crucial factor. where gcd(rα . χn be characters of a group G in a ﬁeld K. . The precise outcome can be determined with the aid of Proposition 5. (ii) By Proposition 5. Definition. 6. . A group homomorphism χ : G −→ K × is called a character of G with values in K. .10.ζ (X) ∈ K[X] are primitive roots of 1. .14. Definition. We have the following splitting ﬁelds and Galois groups. Then χσ : E × −→ E × is a character. Example. .
If χ1 . When p = char K > 0. (i) By Proposition 5. X 4 − 1 is separable over F3 (T ) and has irreducible factors (X − 1). this Galois group only depends on the largest subﬁeld of K which is algebraic over Fp .Proof. .3. Proof. . n) = 1. χn are linearly independent if for t1 . (X + 1) and (X 2 + 1). Artin’s Theorem on linear independence of characters Let G be a group and K a ﬁeld. t1 χ1 (g) + · · · + tn χn (g) = 0. = = with generator σ satisfying σ(ζ) = ζ −1 = −ζ. the functional equation means that for all g ∈ G. . So X n − 1 splits over K(ζ) and each element α ∈ Gal(K(ζ)/K) has the action α(ζ) = ζ rα . hence the roots of minpolyK. 6. Remark.11. χα : K × −→ K × is a character of K × in K.
Let χ1 . 6. χn be distinct characters of a group G in a ﬁeld K. For the inductive assumption. . χk+1 be a set of k + 1 distinct characters for which there are t1 . .
If one of the ti is zero. then χ1 . . Theorem (Artin’s Theorem). αn be the distinct elements of Gal(E/K). . . χk+1 is linearly independent.15. Hence all of the ti must be nonzero. there must be an element g0 ∈ G for which χ1 (g0 ) ̸= χ2 (g0 ). Hence the associated Klinear transformation TrE/K : E −→ K. Then χ1 . Suppose that α1 . 6. we see that t1 χ1 (g0 )χ1 + · · · + tk+1 χk+1 (g0 )χk+1 = 0. . So for all g ∈ G. . . Hence χ2 . contradicting the inductive assumption. The function TrE/K : E −→ K is called the trace mapping of E/K. . . Hence. since the list γα1 . . .16. which demonstrates the inductive step. . So χ1 . . . . .
89
TrE/K (x) = α1 (x) + · · · + αn (x)
. .16. The rest of the statement follows directly from Corollary 6. . . Then there is a z ∈ K for which t1 α1 (z) + · · · + tn αn (z) ̸= 0. in which the coeﬃcient t2 (χ2 (g0 )−χ1 (g0 )) is not zero. not all of which are 0. again contradicting the inductive assumption. . γαn is the same as α1 . suppose that it holds for any n k. .2) by χ1 (g0 ) and subtracting gives t2 (χ2 (g0 ) − χ1 (g0 ))χ2 + t3 (χ3 (g0 ) − χ1 (g0 ))χ3 + · · · + tk+1 (χk+1 (g0 ) − χ1 (g0 ))χk+1 = 0. We proceed by induction on n. . Multiplying Equation (6.17. . . tn ∈ K be a sequence of elements. χn are linearly independent. . α1 (x) + · · · + αn (x) ∈ E Gal(E/K) = K. is surjective. . χr−1 . . . . . . . . . For n = 1 the result is easily veriﬁed. Proof. αn apart from its order. γ(α1 (x) + · · · + αn (x)) = γα1 (x) + · · · + γαn (x) = α1 (x) + · · · + αn (x). .2) applied to g0 g yields t1 χ1 (g0 g) + · · · + tk+1 χk+1 (g0 g) = 0. χr+1 . Let χ1 . Then the function α1 + · · · + αn : E −→ E is a nontrivial Klinear transformation whose image is contained in K. . . χk+1 is linearly dependent. Let t1 . First note that for x ∈ E and γ ∈ Gal(E/K). Corollary. χk+1 is linearly dependent. . αn are distinct automorphisms of the ﬁeld K. Hence the Klinear transformation t1 α1 + · · · + tn αn : K −→ K is nontrivial. . . . . . As χ1 ̸= χ2 . . say tr = 0. Proof.6. . Equation (6. . tk+1 ∈ K not all zero and such that (6.2) t1 χ1 + · · · + tk+1 χk+1 = 0. . . . Let E/K be a ﬁnite Galois extension of degree n and let α1 . Corollary. . . and therefore since χj (g0 g) = χj (g0 )χj (g).
. Proof. 1. then there is a v ∈ E × such that u = vσ(v)−1 . ker NE/K . The characters σ k : E × −→ E × with k = 0. hence in uσ(u) · · · σ n−1 (u) ∈ E ⟨σ⟩ = K.Suppose that E/K is a ﬁnite Galois extension with cyclic Galois group Gal(E/K) = ⟨σ⟩ of order n. the element uσ(u) · · · σ n−1 (u) ∈ E satisﬁes σ(uσ(u) · · · σ n−1 (u)) = σ(u) · · · σ n−1 (u)σ n (u) = σ(u) · · · σ n−1 (u)u. NE/K (u) = uσ(u) · · · σ n−1 (u).33 we introduce a more general notion of radical extension.
90
. Then im δE/K = ker NE/K . So im δE/K Proposition 5. For each u ∈ E × . We call these simple radical extensions and later in Deﬁnition 6. This cannot be identically zero. 6. Explicitly. .18. Proposition. Let u ∈ ker NE/K . so for some w ∈ E.
6. since uσ(u)σ 2 (u) · · · σ n−1 (u)σ n (w) = w.15. Consider the function id +uσ + uσ(u)σ 2 + · · · + uσ(u) · · · σ n−2 (u)σ n−1 : E × −→ E.27. Thus we have u = vσ(v)−1 as required. since σ n (u) = u.19. Theorem (Hilbert’s Theorem 90). where char K n. ζ). where u is a root of f (X) and ζ is a primitive nth root of 1. There is another homomorphism δE/K : E × −→ E × . Notice that uσ(v) = uσ(w) + uσ(u)σ 2 (w) + uσ(u)σ 2 (u)σ 3 (w) + · · · + uσ(u)σ 2 (u) · · · σ n−1 (u)σ n (w) = v. Notice that for u ∈ E × . n−1 are distinct and linearly independent by Artin’s Theorem 6. Then the splitting ﬁeld of f (X) over K has the form K(u. 6. Now using this we deﬁne a group homomorphism NE/K : E × −→ K × . if u ∈ E × and u σ(u) · · · σ n−1 (u) = 1. Let E/K be a ﬁnite Galois extension with cyclic Galois group Gal(E/K) = ⟨σ⟩ of order n.4. NE/K (δE/K (u)) = (uσ(u)−1 )(σ(u)σ 2 (u)−1 · · · σ n−1 (u)σ n (u)−1 ) = 1. Simple radical extensions In this section we will investigate splitting ﬁelds of polynomials of the form X n − a. .3. Our next result is an important generalization of δE/K (u) = uσ(u)−1 . . Let f (X) = X n − a ∈ K[X] be irreducible and separable over K.
NE/K is called the norm mapping for E/K and generalizes the norm mapping for ﬁnite ﬁelds of Section 5. the element v = w + uσ(w) + uσ(u)σ 2 (w) + · · · + uσ(u) · · · σ n−2 (u)σ n−1 (w) is nonzero.
ak ∈ K. L/K is an (iterated ) nKummer extension if L = K(u1 . Gal(K(u)/K) is cyclic since every subgroup of a cyclic group is cyclic. 1 k Note that in this deﬁnition we do not require the polynomials X n − aj ∈ K[X] to be irreducible. . u)/K(ζ)) ▹ Gal(K(ζ. . . n q)/Q(ζn ) is a simple nKummer extension. The Galois group Gal(K(u)/K) is cyclic of order n with a generator σ for which σ(u) = ζu. .23. Let K(u)/K be a simple nKummer extension. Example. . Then in K[X] we have X n − a = (X − u)(X − ζu) · · · (X − ζ n−1 u). . As ⟨ζ⟩ is cyclic of order n.6. In fact. βα(u) = β(ζ rα u) = ζ rα β(u) = ζ rα ζ rβ u = ζ rα +rβ u. .20. u)/K(ζ)) = is abelian. Let n 1 and q ∈ Q. ζ say. u)/K). n − 1. Suppose that un = a ∈ K. Corollary. . Then Q(ζn . u)/K))/ Gal(K(ζ. u)/K)
7
▹
K(ζ)
f }
Gal(K(ζ.22. .21. u)/K(ζ)) is cyclic and Gal(K(ζ)/K)) ∼ Gal(K(ζ. un = ak for some elements a1 . For each α ∈ Gal(K(u)/K) we have α(u) = ζ rα u for some rα = 0. In the more general situation of Proposition 6. Lagrange’s Theorem implies that the image of ρ has order dividing n. Then K(u)/K is a Galois extension and Gal(K(u)/K) is cyclic with order dividing n. then the splitting ﬁeld of f (X) = X n − a over K has the form K(u). ζ. ρ is injective. Then L/K is a simple nKummer extension if L = K(u) where un = a for some a ∈ K. . Clearly the roots of X n − a are distinct and so K(u)/K is separable over K. . . . 1 . 6. Proposition.
is a group homomorphism. K(ζ. u)
a
Gal(K(ζ.
91
. . If K contains a primitive nth root of 1. This means that K(u)/K is Galois. u)/K(ζ))
8 '
K
{id}
6. and so rβα = rα + rβ . where Gal(K(ζ. where u is a root of f (X). The Galois Correspondence identiﬁes the following towers of subﬁelds and subgroups. ρ(α) = ζ rα . Definition. {id} ▹ Gal(K(ζ. Proof. hence  Gal(K(u)/K) divides n. Hence the function ρ : Gal(K(u)/K) −→ ⟨ζ⟩ . Notice that for β ∈ Gal(K(u)/K). √ 6. uk ) where un = a1 . . K(u) is a splitting ﬁeld of X n − a over K. in fact. Since every element of Gal(K(u)/K) is determined by its eﬀect on u. Let K be a ﬁeld with char K n and which contains a primitive nth root of 1.19.
If E/K is a ﬁnite Galois extension with cyclic Galois group of order n.25.18. √ The corresponding group homomorphism ρ : Gal(Q(i)( 2)/Q(i)) −→ ⟨i⟩ has image im ρ = {1. . hence an ∈ K since it is ﬁxed by σ.26. n − 1 are distinct. so by Hilbert’s Theorem 6. Proposition. √ Proof. there is an element a ∈ E for which ζ −1 = aσ(a)−1 . so they must be the n conjugates of a. whence K(a) = E. If each composition factor is a cyclic group of prime order the subnormal series is called a composition series. Then σ(a) = ζa and the elements σ k (a) = ζ k a for k = 0.
92
. Solvability and radical extensions We begin by recalling some ideas about groups. . Proof. 6.√2 (X). A group which is not solvable is called insolvable.5. Hence E/K is a simple nKummer extension. . Since K(a) n = [K(a) : K] and therefore [K(a) : K] = [E : K] = n. Suppose that char K n and there is an element ζ ∈ K which is a primitive nth root of unity. Example.√ √ 6. then there is an element a ∈ E such that E = K(a) and a is a root of a polynomial of the form X n − b with b ∈ K. 6. 5] for further details. Also note that X n − an = (X − a)(X − ζa) · · · (X − ζ n−1 a) = (X − a)(X − σ(a)) · · · (X − σ n−1 (a)). . see [3. this shows that [E : K] = n ⟨i⟩ . 2)/Q(i) is a simple 4Kummer extension with Gal(Q(i. We have ( 2)4 − 4 = 0. A group G is solvable. We have NE/K (ζ −1 ) = ζ −n = 1. Q(i.
in which Gk+1 ▹ Gk and each composition factor Gk /Gk+1 is abelian. 2)/Q(i)) cyclic of order 2. −1} Here is a converse to Proposition 6. Definition. soluble or soluable if there is a chain of subgroups (called a subnormal series) {1} = Gℓ Gℓ−1 ··· G1 G0 = G E. 1. but X 4 − 4 = (X 2 − 2)(X 2 + 2).22.24. we usually write {1} = Gℓ ▹ Gℓ−1 ▹ · · · ▹ G1 ▹ G0 = G. and X 2 − 2 = minpolyQ(i). 6.
6. This follows from the fact that if n quotient group Sn /An ∼ Z/2. so beware when reading other sources! 6. Thus every element of L is expressible in terms of iterated roots of elements
We will need the following Lemma and its Corollary. (ii) If N ▹ G and G/N are solvable then so is G.31.29. =
Now we explain how this relates to ﬁelds and their extensions. a2 . 6. if L/K is a radical Galois extension then so is E/K. . only their order varying for diﬀerent composition series. Remark. Then G is solvable. Proposition. Example. Let E/L be an extension where E is a splitting ﬁeld for the polynomial minpolyK. ak−1 ) k for some dk of K. It is a standard result that we can always reﬁne (i. An is a simple group and An ▹ Sn with
Proof. (i) If G is solvable then every subgroup H G and every quotient group G/N is solvable.e. 6. we assume also that char K = 0. Example. Suppose that ud = a ∈ L with a ̸= 0. Let G be a group. For n 5. 6.. .32.28. Then E/L is a radical Galois extension. (iii) G has a composition series in which one of the terms is a nonabelian simple group.34.30. . an ) with adk ∈ K(a1 . . L/K is a radical extension of K if it has the form L = K(a1 . Recall that a group is simple if it has no nontrivial proper normal subgroups. Let G be a ﬁnite group. Let L/K be a ﬁnite Galois extensions and let L(u)/L be a radical extension. The primes appearing as well as the number of times each occurs are all determined by G. 5. Example.u (X) ∈ L[X]. Definition. so there are d distinct dth roots of unity in E.27. Then G is solvable. . In fact. Then G is insolvable if any of the following conditions holds: (i) G contains a subgroup which is a nonabelian simple group. 1. . and if v is any other root then (v/u)d = 1. several text books make subtle errors or omissions related to this result. Let G be a ﬁnite abelian group. Let G be a ﬁnite pgroup. add extra terms) a subnormal series of a solvable group to obtain a composition series. (ii) G has a quotient group which is a nonabelian simple group. In particular. For simplicity. . Hence there is a primitive dth root of unity ζ ∈ E and the subﬁeld L(ζ. Proof. u) E is normal over
93
. Lemma.6. Proposition. there is always a normal subgroup of a pgroup with index p. for a ﬁnite pgroup G. 6. so in this case we can assume each quotient Gk /Gk+1 is cyclic of order p.33. According to [4]. 6. where p is a prime. . the alternating and symmetric groups An and Sn are insolvable. In the opposite direction we can sometimes see that a group is insolvable. Then X d − a has a d distinct roots in E. a2 . Let K be a ﬁeld and L/K a ﬁnite extension.
so Gal(E radical extension. L′ where L′ /K is a ﬁnite radical Galois extension
6. . But this is clearly a radical extension of L. . ur )/K(ζ. .33. where ζ d = 1. . . . ur−1 )). Definition. Gal(E ′ /K)/ Gal(E ′ /E) ∼ Gal(E/K) which is = ′ /K) is solvable by Proposition 6.7. In the next deﬁnition. . Hence Gal(E ′ /K) is solvable. . . m with d1 · · · dm  d. . . u)/K need not be Galois. which is a radical Galois extension of K. Corollary.35. If L/K is a radical extension then it is contained in a radical Galois extension L′ /K. If L is the splitting ﬁeld of a polynomial f (X) ∈ K[X]. .30. L/K is solvable if L of K. . If r 1 Gr ▹ Gal(E ′ /K) and (E ′ )Gr = K(ζ. We will now show that E ′ /K is a solvable. Then E/K is solvable if and only if the group Gal(E/K) is solvable. then {1} = Gm ▹ Gm−1 ▹ · · · ▹ G0 ▹ Gal(E ′ /K) and Gr−1 /Gr ∼ Gal(K(ζ. this is proved by induction on n using Lemma 6. . . . so E ′ contains a primitive nth root of unity ζ and therefore it contains a primitive dth root of unity for every divisor d of n. then E = L(ζ. Let E ′ be the splitting ﬁeld of X n − 1 over E. the word Galois is superﬂuous because of the preceding results. . .u (X) in E. Theorem. . .30. an ). 6. and since Gal(E/K) is a quotient group of Gal(E ′ /K).L. Proof. u1 . Definition. an .36. ut ). ut ). ut ∈ E are the distinct roots of minpolyK. . . u1 . . = which is abelian by Proposition 6. . u1 . . 6.
94
. ur ). u1 . . . . 6. Now suppose that Gal(E/K) is solvable and let n =  Gal(E/K). . . .22. u. u1 . u1 . Suppose that E E ′ where E ′ /K is a ﬁnite radical Galois extension. . a2 . . . an ) as in Deﬁnition 6. But L(ζ. so E ′ = K(ζ. um ). However. . then f (X) is solvable by radicals over K if L is contained in a radical (Galois) extension of K. is also solvable by Proposition 6. . . . . ur−1 ) for r = 2. . so L(ζ. . . .34. with (E ′ )G0 = K(ζ). u. if u = u1 . . ζ. . u1 . If L/K is a radical Galois extension. . Now Gal(E ′ /E) ▹ Gal(E ′ /K) and by Theorem 6. . u)/L is a radical Galois extension. Gal(E ′ /E) is abelian. Proof. then E = L(a1 . say L = K(a1 . Also.38. .37. Let E/K be a ﬁnite Galois extension. Writing L = K(a1 . ud1 ∈ K(ζ) and udr ∈ K(ζ.
3 2)A3
i
2
A3 = Gal(Q(ζ3 . 2 ζ3 . It is also interesting to reverse the question and ask whether there are extensions which are not solvable.
There are two turning points. 3 2) which we list in the order √ √ √ 2 3 3 3 2.8(iii). Clearly Q(ζ3 .40. 2) = Q(ζ3 )( 2). √ Q(ζ3 .38 with p = 7. Let E C be the splitting ﬁeld of f (X) over Q. √ 6. 3 2) is a radical extension of Q and √ √ 3 3 Q(ζ3 . where we identify each element of the Galois group with = √ a permutation of the three roots of X 3 − 2 in Q(ζ3 .25. We have the following towers of subﬁelds and subgroups related under the Galois Correspondence.20.
95
. Proof. The extension (E ′ )G1 /K(ζ) is radical by Proposition 6.30 and 4. Notice that A3 ∼ Z/3 and S3 /A3 ∼ Z/2. 3 2) 5 S3
e
3 2
√ Q(ζ3 ) = Q(ζ3 . The splitting ﬁeld of the polynomial f (X) = X 5 − 35X 4 + 7 ∈ Q[X] is not solvable. 6. Example. we ﬁnd that f (X) is irreducible over Q. 3 2)/Q) ∼ S3 . f ′′ (X) = 20X 4 − 420X 2 = 20X 2 (X − 21). so by Cauchy’s Lemma this group contains an element of order 5. We should therefore expect to look for a polynomial of degree at least 5 to ﬁnd a Galois group for a splitting ﬁeld to be simple or occur as a composition factor of such a Galois group. Example. This was a famous problem pursued for several hundred years. Now observe that f ′ (X) = 5X 4 − 140X 3 = 5X 3 (X − 28). Hence E ′ /K(ζ) is radical. Using the Eisenstein Test 1. We have already studied this extension in Example 3.
5
3
√ 3 2)/Q(ζ3 ))
Q
y )
{id}
Here Q(ζ3 )/Q is itself a Galois extension and A3 ▹ S3 . 3 2)/Q is solvable. Then Gal(E ′ /K(ζ)) ▹ Gal(E ′ /K) is solvable. each extension (E ′ )Gk+1 /(E ′ )Gk is radical. namely a maximum at x = 0 and a minimum at x = 28. By Theorem 4. √ Proof.39.Clearly K(ζ)/K is radical. Let {1} = Gℓ ▹ Gℓ−1 ▹ · · · ▹ G1 ▹ G0 = Gal(E ′ /K(ζ)) be a composition series. The Galois group of the extension Q(ζ3 . = = so we have the following composition series for S3 : {id} ▹ A3 ▹ S3 . Here is an explicit example over Q. 2 ζ3 . Similarly. 5 divides the order of Gal(E/Q). we ﬁrst recall that the smallest nonabelian simple group is A5 which has order 60. as is E ′ /K. Then f (0) = 7 > 0 > f (28) = −4302585. To ﬁnd examples. √ We know that Gal(Q(ζ3 .
Xn ) by ring isomorphisms ﬁxing k and in the latter case it is by ﬁeld automorphisms ﬁxing k. . Xσ(n) ). . . Suppose that H (1 2 · · · n). = 6. this extends to a ring isomorphism σ· : k(X1 . Definition. . . . . This completes the veriﬁcation of Example 6. .41.44. . . Xn )/ Symn (k)) ∼ Sn . Xn ) called elementary symmetric functions. . . u2 . . .26. Xn )[Y ] we have fn (Y ) = Y n − e1 Y n−1 + · · · + (−1)n−1 en−1 Y + (−1)n en = 0. Xn ) = f (Xσ(1) . then the transposition (1 2) ∈ S5 corresponds to this element. . . It is worth remarking that the most extreme version of this occurs when we ask for a Galois group which is simple. Theorem. Then complex conjugation restricts to an element of order 2 in Gal(E/Q) which interchanges the nonreal roots and ﬁxes the others. Symmetric functions Let k be a ﬁeld. Xn ).6. the only elements of S5 of order 5 are 5cycles.40. Xn ) ∈ k(X1 . ∑ ek = Xi1 Xi2 · · · Xik . but apparently not all simple groups are known to occur as Galois groups of extensions of Q or other ﬁnite subextensions of C/Q. . . . . . . . . . . . 6. There has been a great deal of research activity on this question in the past few decades. . . Xn ) = f (Xσ(1) . Xn )/ Symn (k) is a ﬁnite Galois extension for which Gal(k(X1 . . . . . . . .hence there are three real roots of f (X) and two nonreal complex ones. k(X1 . . . Xn ). so ek ∈ Symn (k). . . .
6. . . = Proof.42. . Varying σ we obtain actions of the group Sn on k[X1 . Xn ] and k(X1 . by taking an appropriate power we can assume that there is a 5cycle of the form (1 2 3 4 5) corresponding to an element of Gal(E/Q) which we can view as a subgroup of S5 . . .43. Lemma. Viewed as a function σ· : k[X1 . Xn ). . Here is an example whose Galois group is A5 . . 6. eσ = ek . . . . Xn ] −→ k[X1 . . 6. Xn ] and its ﬁeld of fractions K = k(X1 . then f (X1 . u3 . . Xn ] by σ · f (X1 . . Xn ) −→ k(X1 . . . . Furthermore. . .
i1 <i2 <···<ik
where 1 k n. . Xn ) = f σ (X1 . . . If we list the roots of f (X) as u1 . . . . . Then H = Sn . . . . Xn ] is a ring isomorphism. . The extension k(X1 . . .
96
. hence it is not solvable. . . It is easy to see that for every σ ∈ Sn . . . . Xn ). Example. . . . u2 being the nonreal roots. Xn ] ⊆ k(X1 . The ﬁeld of symmetric functions on n indeterminates is Symn (k) = k(X1 . The Galois group of f (X) = X 5 + 20X + 16 over Q is Gal(Q(f (X))/Q) ∼ = A5 . . . Xn )Sn So if f (X1 . u4 . . . . . Sn and H contains the elements (1 2) and
The proof is left as an exercise. . Xn ) ∈ Symn (k) ⇐⇒ ∀σ ∈ Sn f (X1 . . There are elements of k[X1 . . . Consider the polynomial ring on n indeterminates k[X1 . . Let n 1. Each permutation σ ∈ Sn acts on k[X1 . . The next lemma shows that Gal(E/Q) ∼ S5 . Xσ(n) ). . this is veriﬁed using Proposition 4. . . . u5 with u1 . Working in k the ring k(X1 .
Hence every element of Symn (k) is a rational function in the ei . If K contains a primitive nth root of unity. . So k(X1 .3. . Xn ) is the splitting ﬁeld of fn (Y ) over Symn (k). . we also have n!  Gal(k(X1 . . . Remark. Corollary. . . Let K be a ﬁeld for which char K ̸= 2 and n 1 be odd. Xn )/ Symn (k)). . en ) is Galois of degree n!. . . Xn )/ Symn (k)) permutes the roots of fn (Y ) and is determined by this permutation. . Using this. this proof shows that the extension k(X1 . . . Now Sn Gal(k(X1 . . . . .] 6. . (b) Determine the group (Z/20)× and describe the eﬀect of each of its elements on Q(ζ20 ) and Q(cos(π/10)) under the action described in Theorem 6. Find all values of n in the following ﬁelds: 1 for which φ(n)  4. 6. .2. 6. . . en ) = Symn (k). . .e. [Hint: what do you know about the centre of G? Use this information to produce a quotient group of smaller order than G. . .5. . Xn )/ Symn (k)) Sn  = n!. . . Describe the eﬀect of each element on Q(ζ24 ) and Q(cos(π/12)) under the action described in Theorem 6. Xn ) : Symn (k)] =  Gal(k(X1 . . Q( 2 i). hence [k(X1 . . . Let p > 0 be a prime and G a group of order G = pn for some n 1. . .2.46.
97
. . . . Xn )/ Symn (k)) = Sn . . .2. 6. . Xn ]Sn = k[e1 . . Xn )/ Symn (k)) = n! and therefore Gal(k(X1 . In this question. .
Combining these inequalities we obtain  Gal(k(X1 . .4. .
(a) What can you say about sin(2π/n) and Gal(Q(sin(2π/n))/Q))? (b) Determine sin(π/12) and Gal(Q(sin(π/12))/Q)). . Let n 1.
But as every element of Gal(k(X1 .
Exercises on Chapter 6 6.. work in the cyclotomic ﬁeld Q(ζ5 ) where ζ5 = e2πi/5 . Xn )/ Symn (k) is not solvable. Q( 3 i). . 6. . . . .1. show that then K contains a primitive 2nth root of unity. (a) Describe the Galois group Gal(Q(ζ5 )/Q) and its action on Q(ζ5 ). . Xn )/ Symn (k)). i. Since k(e1 . en ]. If n 5.6. determine which roots of unity lie √ √ √ Q(i). . Q( 5 i). . .45. . . In fact. en ) Symn (k) we can also deduce that k(e1 . . . . the extension k(X1 .hence the roots of this polynomial are the Xi . Xn )/k(e1 . . . 6. k[X1 .
6. (a) Describe the elements of (Z/24)× explicitly and verify that this group is isomorphic to Z/2 × Z/2 × Z/2. Analogous results are true for polynomials. Show by induction on n that there is a normal subgroup N ▹ G with N  = pn−1 .
Prove Lemma 6. . if p ≡ 3 (mod 4).
(c) Conclude that
if p ≡ 3 (mod 4).7.10.9. . (a) Show that the function T : E −→ E. Hence show that √ −1 + 5 cos(2π/5) = . 6.41.] 6. (b) If v ∈ E has TrE/K (v) = 0.
98
. show that there is a w ∈ E such that v = w − σ(w).
if p ≡ 1 (mod 4).8.
ξ= Show that
∏
r −r (ζp − ζp ) ∈ Q(ζp ).(b) Determine the minimal polynomial of cos(2π/5) over Q. Xn ]Sn is deﬁned by ∑ sk = Xik . let p be an odd prime and let ζp = e2πi/p ∈ Q(ζp ) (a) Consider the product
(p−1)/2
C.18. Let E/K be a Galois extension with cyclic Galois group Gal(E/K) = ⟨σ⟩ of order n.
r=1 p−1 ∏ r=1
ξ = (−1) (b) Deduce that p ξ2 = −p ±√p ξ= ±√p i
2
(p−1)/2
r (1 − ζp ). the kth power sum sk ∈ k[X1 . then consider w= 1 (TrE/K t) ( ) vσ(t) + (v + σ(v))σ 2 (t) + · · · + (v + σ(v)σ 2 (t) + · · · + σ n−2 (v))σ n−1 (t)
and adapt the proof of Hilbert’s Theorem 90 in Theorem 6.18. 6. T (u) = u + σ(u) + σ 2 (u) + · · · + σ n−1 (u). [Hint: Show that there is an element t ∈ E for which TrE/K t ̸= 0. In this question. . [Hint: show that every 2cycle of the form (i i + 1) is in H by considering elements of the form (1 2 · · · n)r (1 2)(1 2 · · · n)n−r .
takes values in K and use this to deﬁne a Klinear mapping TrE/K : E −→ K. see Theorem 6. . This question is about an additive version of Hilbert’s Theorem 90. using TrE/K in place of NE/K .
1 i n
Prove the formula sk = e1 sk−1 − e2 sk−2 + · · · + (−1)k−1 ek−1 s1 + (−1)k kek . √ √ and also p ∈ Q(ζp ) if p ≡ 1 (mod 4) and p i ∈ Q(ζp ) if p ≡ 3 (mod 4). if p ≡ 1 (mod 4). (a) For n 1 and 1 k n.] 6. 4 For which other angles θ is cos θ a root of this minimal polynomial? What is the value of sin(2π/5) ? (c) Find the tower of subﬁelds of Q(ζ5 ) and express them as ﬁxed ﬁelds of subgroups of Gal(Q(ζ5 )/Q).
99
..
j1 j2 ··· jk
i. h3 in terms of the elementary symmetric functions e1 .(b) For n
1 and 1
k
n. the sum of all the monomials in the Xi of degree k. express h1 . (i) For large values of n. the total symmetric function is deﬁned by ∑ hk = Xj1 Xj2 · · · Xjk . h2 .e. e 3 . (ii) Show that the power sum functions sk of the previous question satisfy sk = −(h1 sk−1 + h2 sk−2 + · · · + hk−1 s1 ) + khk . e 2 .
.
[7] J. Lidl & H. Lang. Addison Wesley (1993). Escoﬁer. [Highly recommended ] [4] T. Galois Theory. B. [6] R. Chapman and Hall (1989). [Highly recommended. [8] I. A counterexample in Galois theory.]
101
. Algebra. ISBN 0 521 39231 4. Hungerford. [2] JP. Niederreiter. Dover Publications (1998). especially for its historical notes] [3] J. Rotman. ISBN 0 486 62342 4. A First Course in Abstract Algebra. Finite Fields. [Very highly recommended. Galois Theory. ISBN 0 201 33596 4.Bibliography
[1] E. ISBN 0 201 55540 9. 54–57. Artin. W. Stewart. Cambridge University Press (1997). SpringerVerlag (1998). ISBN 0 412 345501. ISBN 0387987657. SpringerVerlag. Fraleigh. Galois Theory. New York (2001). Addison Wesley (1999). [5] S. ISBN 0 387 98541 7. American Mathematical Monthly 97 (1997). Galois theory.
.
Clearly {n ∈ Z : n > 0 and nr = 0 for all r ∈ R} ⊆ {n ∈ Z : n > 0 and n1 = 0}. so P ⊆ Q ∩ R. If 0 < n ∈ Z and n1 = 0. For R = Fp [X].
n n
so {n ∈ Z : n > 0 and n1 = 0} ⊆ {n ∈ Z : n > 0 and nr = 0 for all r ∈ R}. (a) For any subring R ⊆ C. Hence these sets are in fact equal. φ is not surjective. (a) Recall from the Isomorphism Theorems of basic Ring Theory that φ−1 Q ▹ R. v ∈ R with uv ∈ φ−1 Q. v ∈ R and R is an integral domain.5. (b) Q is a ﬁeld and Z ⊆ Q is a subring which is not a ﬁeld. Now by deﬁnition of characteristic. 1. v ∈ S and suppose that uv = 0. 1. 1. then u = 0 or v = 0 since u. Suppose that u. (d) We have P ⊆ Q ∩ R ▹ R with P ▹ R maximal. The main thing to check is that φ(u + v) = φ(u) + φ(v) which is a consequence of the Idiot’s Binomial Theorem. If n > 1 then Matn (A) is not commutative. (b) Consider the inclusion function inc : R −→ S.3. we need to show it is a prime ideal. Consider the unit homomorphisms η : Z −→ R and η ′ : Z −→ S. hence u ∈ φ−1 Q or v ∈ φ−1 Q. Then for n ∈ Z. 1. When char R = p > 0 they must both be nonempty. η ′ (n) = η(n). R is an integral domain with characteristic subring Z and char R = 0. nr = r + · · · + r = (1 + · · · + 1)r = (n1)r = 0r = 0. (b) The characteristic subring of A[X] is the same as that of A and char A[X] = char A. φ is not injective.Solutions
Chapter 1 1. then the characteristic subring of Matn (A) is the same as that of A and char Matn (A) = char A. in any case it always has zerodivisors since any singular matrix is a zerodivisor. In fact Q only needs to be a proper ideal of S for this argument to work.1. so this result follows from (a). (c) Consider Z ⊆ Q. while for R = Fp [X]/(X 2 ). then the zeroideal (0)Q ▹ Q has (0)Q ∩ Z = (0)Z ▹ Z but this is not maximal in Z since for any prime p > 0.2. then φ(u)φ(v) = φ(uv) ∈ Q and so φ(u) ∈ Q or φ(v) ∈ Q. (a) Let u. then for every r ∈ R. (p)Z ▹ Z is a (maximal) ideal that properly contains (0)Z . then inc−1 Q = Q ∩ R.4. (c) If we identify A with the subring of scalar matrices in Matn (A). char R = min{n ∈ Z : n > 0 and n1 = 0} = min{n ∈ Z : n > 0 and nr = 0 for all r ∈ R}.
. so ker η ′ = ker η and therefore char S = char R. A[X] is an integral domain if and only if A is an integral domain.
d (X) = ψa. (u1 +v1 i)(u2 +v2 i) = (u1 u2 −v1 v2 )+(u1 v2 +u2 v1 )i. The only proper ideal of k is the zero ideal (0). (a) Addition and multiplication follow from the obvious formulae (u1 +v1 i)+(u2 +v2 i) = (u1 +u2 )+(v1 +v2 )i.b (f (X)) = deg f (X). 1. Hence every nonzero element of Q[i] has an inverse.cb+d (X). To see that Q[i] is a ﬁeld. (u − vi)(u + vi) = (u + vi)(u − vi) = u2 + v 2 ̸= 0.d = ψca. By the uniqueness part of the Homomorphism Extension Property.cb+d .b (p(X))  g(X)h(X) for g(X). therefore Q[i] is a ﬁeld. (ii) Suppose that ψa. so they are subrings of the ﬁeld C.
−1 Therefore these are inverse isomorphisms.b (k(X))p(X) 104
.b (f (X)) = f (aX + b) = We have ψa. h(X) ∈ k[X]. (a) Existence and uniqueness of such an ψa. we have
−1 −1 −1 ψa.−ba−1 ◦ ψa. so ker φ = (0). ψa−1 .b is an isomorphism.b (p(X)).b (cX + d) = c(aX + b) + d = caX + (cb + d) = ψca.b (h(X)) = ψa.1. u.b ◦ ψc.−ba−1 = id = ψa−1 .8.22 and its eﬀect on f (X) = n ri X i ∈ R[X] where ri ∈ R is i=0 ψa. ψa. this shows that deg ψa. Since cn an ̸= 0. notice that if u + vi ̸= 0 with u. they are both integral domains.b (f (X)) =
n ∑ i=0
ci (aX + b)i
= cn an X n + terms of lower degrees in X.b follow from the Homomorphism Extension ∑ Property 1. v ∈ Z and n ̸= 0. so u v + 2 i ∈ Q(i) 2 +v u + v2 is the inverse of u + vi. If a is a unit then ψa−1 . 1.
ψa.b (a−1 X − ba−1 ) = X.b ◦ ψa−1 . Then ) ( ) ( (u + vi) 1 (u + vi) = inc∗ = (u + vi). Choose k(X) ∈ k[X] so that g(X)h(X) = k(X)ψa. Since ψa.6.1. so by the uniqueness part of the Homomorphism Extension Property. v ∈ Q.−ba−1 : R[X] −→ R[X] has the property that ψa−1 . we have ψa. Now n ∑ i=0
ri (aX + b)i .b .
with Z[i] and Q[i] both closed under these operations and containing 1 = 1 + 0i as a unity.−ba−1 (aX +b) = X and ψa.b (g(X))ψa. 1 (b) & (c) The crucial point is that every element of Q[i] can be written as (u + vi) with n n. inc∗ n n + 0i n u2 so the latter element is in the image of inc∗ which must therefore be a surjection.7. then deg f (X) = n. ∑n (b) (i) If f (X) = i=0 ci X i ∈ k[X] with ci ∈ k and cn ̸= 0.b . 1. by Qu.−ba−1 = ψa.b ◦ψc.
Then a has an inverse in k[[X]] only if there is a b = ∑∞ ℓ k=0 bℓ X ∈ k[[X]] with ab = 1.−1 −1 and as p(X) is prime.
1. X −k0
r=0
so inc∗ is surjective. 1. (iii) This follows from (ii) and Proposition 1. k=0 k=0 k=0 ℓ=0
Clearly k[X] ⊆ k[[X]] is a subring.ℓ0 } j=0
Clearly k[[X]] ⊆ k((X)) is a subring.b (g(X)) or p(X)  ψa. Hence ab ̸= 0.k2 }
(
ak X k )(
∞ ∑ ℓ=ℓ0
bℓ X ℓ ) =
∞ ∑
(
aℓ bk−j )X k . 1. Here f (X) = −X 3 − X 2 + X + 1 and d(X) = −X 3 − X with f (X) = d(X) + (−X − X 2 + 1) = d(X) + (2X 2 + 2X + 1). p(X)  ψa. ∑∞ k (b) Let a = k=0 ak X ∈ k[[X]]. in particular this forces a0 ̸= 0 since otherwise the lowest term in X in ab would be of degree greater than 0.9. b ∈ k[[X]] we may write a=
∞ ∑ k=k0
ak X . 1. Here f (X) = (3X − 3)d(X) + (−9X + 7).
105
ρ(f (X)) = f (X).12. (a) Addition and multiplication are given by the usual formulae
∞ ∞ ∞ ∑ ∑ ∑ k k ak X ) + ( bk X ) = (ak + bk )X k . The reduction modulo p function ρ : Z[X] −→ Fp [X].
.
k
b=
∞ ∑ ℓ0
bℓ X ℓ
with ak0 ̸= 0 ̸= bℓ0 .11.
k ∑
k=k1
k=min{k1 . Given two nonzero elements a. if a0 ̸= 0. So k[[X]] is an integral domain. Conversely. Then the lowest degree term in ab is ak0 bℓ0 X k0 +ℓ0 with ak0 bℓ0 ̸= 0. ∑∞ ak X k ∈ k((X)) with
k=min{k0 .b (p(X))  g(X) or ψa.b (p(X))  h(X) and so ψa. then we can inductively solve the system of equations a0 b0 = 1.b (h(X)).b (p(X)) is prime.
to ensure that ab = 1. ( k=0 k=0 k=0 ∞ ∞ ∞ k ∑ ∑ ∑∑ k k ( ak X )( bk X ) = ( aℓ bk−ℓ )X k .10.30. (c) We can deﬁne make the set k((X)) of all such ﬁnite tailed Laurent series into a ring with addition and multiplication deﬁned by (
∞ ∑
ak X ) + (
∞ ∑ k=k0
k
∞ ∑ k=k2
bk X ) =
k
∞ ∑
(ak + bk )X k .
n ∑ ℓ=0
aℓ bn−ℓ = a0 bn + a1 bn−1 + · · · + an bn = 0 (n
1). Hence ψa. Notice that every element k0 < 0 can be written as ∑ ( ar+k0 X r )X k0 .
r=0
k=k0
The inclusion inc : k[[X]] −→ k((X)) extends to the monomorphism inc∗ : Fr(k[[X]]) −→ k((X)) for which ( ∑∞ ∞ r) ∑ r=0 ar+k0 X inc∗ =( ar+k0 X r )X k0 .
1. The image is The image of ε−√2 √ √ ε√2 Q[X] = Q[ 2] = {a + b 2 : a.14. If f (X) = g(X)h(X) with g(X). so the evaluation homomorphism εω2 has εω2 Q[X] = εω Q[X] and ker εω2 = ker εω .13. So f (X) must be irreducible.15. α i. ker εω = (X 2 + X + 1) ▹ Q[X]. then in R[X]. I3 = (X 2 −2). Then ker ε−α = ker εα i = ker ε−α i = (X 4 − 2) ▹ Q[X] but although ε−α Q[X] = Q[α]. I1 = (X 2 +1). then √ 4 εα R[X] = R[α] = {a + bα : a. b ∈ Q}. So X 3 + X − 1 is irreducible. I5 = (X 2 +2). b. But this is impossible since f (X) is irreducible. we have εα i Q[X] = ε−α i Q[X] = Q[α i] = {a + bα i + cα2 + dα3 i : a. h(X) ∈ Z[X]. If we replace Q by R. √ Let α be a root of X 4 − 2. where deg g(X) < deg f (X) = deg f (X) and deg h(X) < deg f (X) = deg f (X). The other complex roots of X 4 − 2 are −α. √ 1. X 5 − X + 1 is irreducible modulo 3 and 5 so is itself irreducible. I6 = X 2 +X +1. then f (X) = ρ(g(X)h(X)) = ρ(g(X))ρ(h(X)) = g(X)h(X). 1. X 3 − X + 1 reduces modulo 3 to an irreducible since it has no roots modulo 3. X 3 + X − 1 reduces modulo 2 to an irreducible since it has no roots modulo 2. ker εα = (X − 2) ▹ R[X]. I2 = (X 2 +2). d ∈ Q} ̸= Q[α]. so εα i Q[X] ̸= εα Q[X] since one of these is a subset of R but the other is not. We have εα Q[X] = Q[α] = {a + bα + cα2 + dα3 : a. −α i (notice that two of these are real while the other two are not).
where (X 2 + X + 1) ▹ Q[X] is a maximal ideal. b ∈ Q}. √ √ √ √ √ 4 4 X 4 − 2 = (X 2 − 2)(X 2 + 2) = (X − 2)(X + 2)(X 2 + 2). Notice that ω = (−1 + 3i)/2 = ζ3 is a primitive 3rd root of unity and is a root of the irreducible polynomial X 2 + X + 1 ∈ Q[X]. ker εα = (X 4 − 2) ▹ Q[X]. So X 3 − X + 1 is irreducible. b ∈ R} = R. We have ker ε√2 = ker ε−√2 = (X 2 − 2) ▹ Q[X] which is a maximal ideal.
106
. d ∈ Q}. c. X 5 + X − 1 = (X 3 + X 2 − 1)(X 2 − X + 1) and 5X 3 − 10X + X 2 − 2 = (5X + 1)(X 2 − 2) so neither of these is irreducible. √ is ε−√2 Q[X] = Q[ 2] = ε√2 Q[X]. I4 = (X − 2). c. The other complex root of X 2 + X + 1 is ω 2 .16. b. √ 1. Then εω Q[X] = Q[ω] = {a + bω : a. X 3 + 2X + 1 ≡ X 3 − X + 1 (mod 3) so this polynomial reduces modulo 3 to an irreducible and so is irreducible. deg g(X) < deg f (X) and deg h(X) < deg f (X).is a ring homomorphism. If α = 4 2.
and the latter ideal is maximal.
108
so q = −20 and p = 6.
107
. − 3 + 2. while for f (X) we have 5. 3). 2 27q 2 + 4p3 = 108. Work backwards with Cardan’s method. . Thus α is a real root of f (X) = X 3 + 6X − 20. Therefore α = 2. To see that the homomorphism Aﬀ 1 (k) −→ Autk (k[X]). If α2 + 2 = 0.
πi/4 r
1. = . Then φk[X] = {c + 0 + c1 φ(X) + · · · + ck φ(X)k : c0 . take q 27q 2 + 4p3 28 − = 1. . 3 − 1. 1.
√ 2 u = 2e ω = √ (1 + i)ω r = (1 + i)ω r (r = 0. . Suppose that show that n > 1. where X 2 + X + 2 has no real roots.19. b ∈ R} = C. then ε−α R[X] = R[−α] = {a − bα : a. If n = 0 then φk[X] = k ⊆ k[X] so φ would not be surjective.
described in the Proof of Example 1. − 3 − 1. Let φ(X) = a0 + a1 X + · · · + an X n with ai ∈ k and an ̸= 0. For β. hence we must have n 1. 2 √ √ √ √ For the roots of g(X) we obtain 2. suppose that φ ∈ Autk (k[X]) is any automorphism.√ Similarly.17. 1. if α = − 4 2.60 is surjective. A −→ αA−1 . then α ∈ R and / εα R[X] = R[α] = {a + bα : a. Therefore β = 1.18. take q − = 10. 2 108 27 3 + X − 2 for which 1 is also a root and so q = −2 and p = 1. . 1. c1 . which has roots Thus we can take √ −2 ± 2i = ( 2)3 e3πi/4 . √ ker ε−α = (X +
√ 4 2) ▹ R[X]. Using Cardan’s method we have to solve the quadratic equation U 2 + 4U + 8 = 0. where X 2 + 2X + 10 has no real roots. 3 + 2. First change variable to obtain g(X) = f (X + 3) = X 3 − 6X + 4. Thus β is a real root of g(X) = X g(X) = (X − 1)(X 2 + X + 2). b ∈ R} = R.
ker εα = (X 2 + 2) ▹ R[X]. ck ∈ k} = k[X]. For α. Notice that 2 is a real root of this polynomial and f (X) = (X − 2)(X 2 + 2X + 10).
so the roots of Φpk (X) are their pk−1 st roots which are the primitive pk th roots of 1. ∏
0 j k
k−1 (X p ). Calculation. k k−1 k−1 k−1
)p − 1 = (X p
k−1
− 1)Φp (X p
). (b) Using the formula of Equation 1. / which gives a contradiction.
where cr ≡ 0 (mod p) and c0 = p. As these are monic polynomials of
108
.
k−1 k−1
− 1 ≡ (X − 1)p
k−1
(mod p)
k−1
+ c′ (X − 1)(p−2)p p−2
+ · · · + c′ (X − 1)p 1
k−1
+ c′ .20. The Idiot’s Binomial Theorem gives Xp so Φpk (X) = (X − 1)(p−1)p where c′ ≡ 0 (mod p).21. we have Φpk (X) = Φp (X p
k−1
) = (X p
k−1
− 1)p−1 + cp−2 (X p
k−1
− 1)p−2 + · · · + c1 (X p
k−1
− 1) + c0 .22. r c′ = Φpk (1) = Φp (1) = c0 = p. Therefore φ(X) = a0 +a1 X and so φ = αA for some A ∈ Aﬀ 1 (k). This shows that Φn (X)  Φp1 ···pk (X the same degree they are equal. 0
and deg Φp1 ···pk (X p1
r −1 r1 −1 ···pkk
r r ) = φ(p1 · · · pk )pr1 −1 · · · prk −1 = (p1 − 1) · · · (pk − 1)p11 −1 · · · pkk −1 . 1. 0 so the Eisenstein Test can be applied to show that Φpk (X) is irreducible over Q. Also.But if k > 0 and ck ̸= 0 then deg(c + 0 + c1 φ(X) + · · · + ck φ(X)k ) = kn > 1. (c) First notice that
r r deg Φn (X) = φ(n) = (p1 − 1) · · · (pk − 1)p11 −1 · · · pkk −1 . 1. so X ∈ φk[X]. hence (ξ p1
) is a root of
Φp1 ···pk (X). since cyclotomic polynomials are irreducible. (a) We have X p − 1 = (X p so by (1.5). In fact.4.
). we ﬁnd that Φ20 (X)  (X 2 + 1)(X 8 − X 6 + X 4 − X 2 + 1). So we must have n = 1. we must have Φ20 (X) = X 8 − X 6 + X 4 − X 2 + 1. each root ξ of Φn (X).
Φpj (X) = Φp (X p
)
∏
0 j k−1
Φpj (X).
r −1 r1 −1 ···pkk r −1 r −1 p11 ···pkk
(ξ p1 and no smaller power of (ξ p1
r1 −1
r −1 r1 −1 ···pkk
···pkk
r −1
) has this property. 1.
and therefore Φpk (X) = Φp The complex roots of Φp (X) are the primitive pth roots of 1. 1 k
r1 −1
so deg Φn (X) = deg Φp1 ···pk (X p1
···pkk
r −1
). We have deg Φ20 (X) = φ(20) = φ(4)φ(5) = 2 × 4 = 8 and X 20 − 1 = (X 10 − 1)(X 10 + 1) = (X 10 − 1)(X 2 + 1)(X 8 − X 6 + X 4 − X 2 + 1). )p1 ···pk = ξ n = 1. Since the roots of X 10 − 1 are the 10th roots of unity.
. ζ5 = ζ5 . (a) In K[X].n−1 gcd(t.. the only root of this polynomial in K must be 1. 1. −2 −1 2 ζ5 + ζ5 = (ζ5 + ζ5 )2 − 2 = 4 cos2 (2π/5) − 2.. we have −2 −1 2 (ζ5 + ζ5 ) + (ζ5 + ζ5 ) + 1 = 0.n−1 gcd(t. Since 2  φ(n) when n > 2 and the result is immediate when n = 2... we see that desired equation always holds.n)=1 t (ζn − X)
t=1...n−1 gcd(t.. so
t=1. −1 −2 4 3 Rearranging and using the formulae ζ5 = ζ5 .
We also have Φ5 (X) = X 4 + X 3 + X 2 + X + 1.n)=1
= X −φ(n) = X −φ(n) = X −φ(n) = X −φ(n)
∏ ∏
t (1 − Xζn )
t=1...n−1 gcd(t. We have
−1 ζn + ζn = e2πi/n + e−2πi/n
= (cos(2πi/n) + sin(2πi/n) i) + (cos(2πi/n) − sin(2πi/n) i) = 2 cos(2πi/n).10..n−1 gcd(t. by the Idiot’s Binomial Theorem 1.n)=1
= (−1)φ(n) X −φ(n)
∏
t (X − ζn )
t=1.n)=1
= (−1)φ(n) X −φ(n) Φn (X).43.n)=1
Φn (X −1 ) =
∏
t (X −1 − ζn )
t=1.
hence 4 cos2 (2π/5) + 2 cos(2π/5) − 1 = 0. X p − 1 = X p + (−1)p = (X + (−1))p = (X − 1)p ... m m X np − 1 = (X n − 1)p
109
....1.... Now we have
−1 ζ5 + ζ5 = 2 cos(2π/5)...25...24. By the Unique Factorization Property 1. Similarly..n−1 gcd(t. Φn (X) =
∏
t (X − ζn ). Thus a suitable polynomial is 4X 2 + 2X − 1 ∈ Q[X]..n−1 gcd(t. 1. so
4 3 2 ζ5 + ζ5 + ζ5 + ζ5 + 1 = 0...33..23.n)=1 n−t (1 − Xζn )
t=1. By Theorem 1.n)=1
∏ ∏
−t (1 − Xζn )
t=1.
1. q) : Q( p)] = 1 then q ∈ Q( p). we have [Q(i) : Q] 2. / 2. b2 ) = 1. By Theorem 2. It is obvious that [Q( p. √ √ √ √ √ √ √ √ 2. we obtain 1 2 1 2 (b′ )2 p2 q = (b′ )2 q 2 p. i is a basis for Q(i) over Q. contradicting the fact 1 2 √ √ that gcd(b1 . so v ̸= ±u. y ∈ R. so 1. v). i) : Q( 3)] = 2. b2 ) = 1. if [Q( p. Notice that if v = ±u then b = v 2 = u2 = a which is impossible.
√ √ √ If b = 0 then q ∈ Q which contradicts the result of Qu. 2 1 From this we obtain p  b′ and q  b′ . then x + yi = 0 ⇐⇒ x = y = 0.4. Chapter 2 2. Writing b1 = b′ q and b2 = b′ q for suitable b′ .6.
2. 2 1 hence (b′ )2 p = (b′ )2 q. First notice that [Q( 3) : Q] = 2 (with Qbasis 1. q) : Q( p)] 2. i ∈ Q( 3) / √ √ √ √ and since i2 + 1 = 0. i) = Q( 3)(i) has [Q( 3. but then p  b1 as well as p  b2 . i) : Q] = [Q( 3. 2. i) : Q( 3)] [Q( 3) : Q] = 2 × 2 = 4. Q( 3. Since 1. b2 ∈ Z and gcd(b1 . i span the Qvector space Q(i).
110
. Also. As in (a) we have X p − a = X p − up = (X − u)p . √ √ say q = a + b p for some a. √ √ √ √ 2. 3) and Q( 3) R. giving the simultaneous pair of equations a2 + b2 p = q. we obtain b2 q = b2 p 2 1 and so p  b2 and q  b1 .2. b′ ∈ Z. Writing b = b1 /b2 with b1 .1. b ∈ Q. If a = 0 then q = b p.5.and the only roots of this must be nth roots of 1. But also if x. (b) If u ∈ K is a root of this polynomial then up = a. so u is the only root in K. 2ab = 0. Then √ √ q = (a + b p)2 = (a2 + b2 p) + 2ab p. 2 (u − v)(u + v) u2 − v 2 a−b = = ∈ K(u + v).3. v) and therefore K(u + v) = K(u.6(ii). √ √ √ √ [Q( 3. So q ∈ Q( p). Then u−v = Hence u= So K(u.4. 2. u+v u+v u+v
K(u. v) 1 ((u + v) + (u − v)) ∈ K(u + v). This is similar to Example 2. 2 K(u + v) v= 1 ((u + v) − (u − v)) ∈ K(u + v). Arrange the induction carefully. Hence [Q(i) : Q] = 2.
sin(2π/5) i = ζ5 − ζ5 ∈ Q(ζ5 ). so T5 (cos 2kπ/5) − 1 = 0. Hence we have ) ) 1( 1( −1 −1 cos(2π/5) = ζ5 + ζ5 ∈ Q(ζ5 ). 2. so α(21/n ) ∈ En is also a real nth root of 1. 2 2 (c) This can be found by repeated use of the double angle formula
cos(A + B) = cos A cos B − sin A sin B. For k = 1. 4. therefore Q(cos(2π/5)) ∼ Q[X]/(4X 2 + 2X − 1). cos(5(2kπ/5)) = cos(2kπ) = 1. 2. = (e) Q(ζ5 )
2
[Q(cos(2kπ/5)) : Q] = 2.6. C
2
R
∞
∞
√ Q( 3)
tt tt tt t tt
2
√ Q( 3. so Lr ∩ Ls = Q whenever r ̸= s. If n is even. But also ζ5 ∈ Q(ζ5 ) and ζ5 = cos(2π/5) − sin(2π/5) i ∈ Q(ζ5 ). 1. i)
2
uu uu 2 uu uu u
uu uu uu u 2 uuu u
Q(i)
2
Q
rr rr rr r rr 2 rr
√ Q( 3 i)
2. L3 = Q( 3 i). (d) For k = 0.
Q(cos(2π/5))
2
Q 2. 2. 3.8. The only real subﬁeld amongst these is L1 . cos(2kπ/5) is a root of 4X 2 + 2X − 1. 4. see Remark 6. The polynomial Tn (X) expressing cos nθ in terms of cos θ is called the nth Chebyshev polynomial. We can realize this automorphism starting with the evaluation
111
. So each of the numbers cos(2kπ/5) is a root of the polynomial T5 (X) − 1 = (X − 1)(4X 2 + 2X − 1)2 . the only possibility is α(21/n ) = 21/n . (a) Since 5 is a prime. so α = id.7.
−1 −1 (b) We have ζ5 = cos(2π/5) + sin(2π/5) i ∈ Q(ζ5 ). This is similar to the previous question. the possibilities are α(21/n ) = ±21/n . L2 = Q(i). Then [Lr ∩ Ls : Q] > 1 ⇐⇒ Lr ∩ Ls = Lr = Ls . i) are distinct and are extensions of Q having degree 2: √ √ L1 = Q( 3). 3.9. (a) If α ∈ AutQ (En ) then α(21/n )n = α(2) = 2.√ The following three subﬁelds of Q( 3. If n is odd. [Q(ζ5 ) : Q] = [Q[X]/(Φ5 (X)) : Q] = φ(5) = 5 − 1 = 4.
±e these are √ √ √ √ 3 1 3 1 3 1 3 1 + i. so α = id. 3.e. − i.
{ {{ {{ { {{
3
E12
2
E4
2
{ {{ {{ { {{ 3 gg gg gg 2 gg
E6
gg gg 2 gg gg { {{ {{ { {{ 3
E2
E3
Q
(d) This element is a root of the polynomial (X − (21/2 + 21/3 ))(X − (−21/2 + 21/3 )) = X 2 − 2(21/3 )X + 22/3 − 2 ∈ E3 [X]. Chapter 3 3. Writing ω = e2πi/3 .9. So 21/2 + 21/3 is in E6 and / E12 and none of the others. √ 1 ± 3i p1 (X) = X 4 − X 2 + 1: The polynomial X 2 − X + 1 has the complex roots e±πi/3 = . By Theorem 2.1. / 1/2 + 21/3 ∈ E . If 21/2 + 21/3 ∈ E3 then 21/2 ∈ E3 . so it cannot lie in E4 since 21/2 + 21/3 ∈ E2 and 3 [E4 : E2 ] = 2. (c) Assuming there are only 6 such subﬁelds. which would imply 2 = [E2 : Q]  [E3 : Q] = 3 which is false. so 21/2 + 21/3 ∈ E3 . − + i. so 2 ±πi/6 . they form the following tower. Explicitly the four roots of p1 (X) are the complex square roots of these numbers. 21/2 + 21/3 is a root of a similar argument shows that 2 / 2 (X − (21/2 + 21/3 ))(X − (21/2 + 21/3 ω))(X − (21/2 + 21/3 ω 2 )) = X 3 − 3(21/2 )X 2 + 6X − (2 + 2(21/2 )) ∈ E2 [X].2. − − i. so it is certainly an element of E6 which is the only degree 2 extension of E3 . On passing to the quotient homomorphism of 2 ε′ 1/n we obtain an automorphism τn of En under which τn (21/n ) = −21/n . Clearly. 1/n α(2 ) = ±21/n if n is even. 2 2 2 2 2 2 2 2
112
. Hence these three conditions are indeed equivalent. t is algebraic over K if and only if ker εt ̸= (0).. i. This contradiction shows that α(21/n ) = 21/n for every n. If for some n we have α(21/n ) = −21/n then −21/n = α(21/n ) = α(21/2n )2 > 0 since α(21/2n ) ∈ R. 2 (b) Since E R.homomorphism ε21/n : Q[X] −→ En and precomposing with the isomorphism ψ : Q[X] −→ Q[X] for which ψ(X) = −X to form ε′ 1/n = ε21/n ◦ ψ. The diagrams at the bottom indicate useful subﬁelds of the splitting ﬁelds occurring in each of these examples. (ii) ⇐⇒ (iii).e. (i) ⇐⇒ (ii).. an automorphism α ∈ AutQ (E) has the eﬀect 21/n if n is odd. i.
ζ3 )) ∼ S3 . so can be identiﬁed with the unique permutation σα ∈ S3 for which α(ui ) = uσα (i) (i = 1. 5. − √ − √ i. 1. 2. − i. therefore AutQ (Q( 3 2. √ √ √ √ √ √ 3. 3. σα1 = (1 2 3). − − i. √ − √ i. We ﬁnd that (using cycle notation for permutations) σα′ = (1 2). List the three roots of X 3 −2 as u1 = 3 2. then using the Idiot’s Binomial Theorem we have (X − t)p = X p − tp = X p − T.√ The splitting ﬁeld is E = Q( 3. 10) : Q] = 4 and the element 5 + 10 has degree 4 with √ √ minimal polynomial X 4 − 30X 2 + 25 which has roots ± 5 ± 10. √ p3 (X) = X 4 + 2: The roots are the four 4th roots of −2.48.. This also gives the factorization of g(X) into linear factors over k(T ). 3)
6
√ Q( 4 2. + i. so the splitting ﬁeld of p4 (X) over Q is the same as that of Φ5 (Y ) over Q and this is the cyclotomic ﬁeld Q(ζ5 ) where ζ5 = cos(2π/5) + sin(2π/5)i with [Q(ζ5 ) : Q] = 4. σα2 = (1 3 2). 2.e. these are √ √ √ √ √ √ √ √ √ √ √ √ 6 6 6 6 6 6 6 6 √ √ 2 2 3 2 2 3 2 2 3 2 2 3 6 6 2. i. 2. 3. Explicitly.4. − √ + √ i. 2
3.3. sin(2π/5)i)
2
√ Q( 3)
2
√ Q( 3)
2
√ Q( 4 2)
4
Q(cos(2π/5))
2
Q
Q
Q
Q
√ √ √ 2 3. 2 2 2 2 2 2 2 2 √ √ √ √ The splitting ﬁeld is E = Q( 6 2. 4. 3). σα′ = (1 3). 4 2e(2k+1)πi/4 for k = 0. sin(2π/5)i). 10)/Q: Here [Q( 5. i)
2
Q(cos(2π/5). hence it is the only root of g(X) in k(T ). p4 (X) = X 4 + 5X 3 + 10X 2 + 10X + 5: Notice that p4 (Y − 1) = Y 4 + Y 3 + Y 2 + Y + 1 = Φ5 (Y ). √ √ Q( 6 2. i. 1 √ These are the six elements of S3 . i) and [E : Q] = 8. 4 4 4 4 4 4 4 4 2 2 2 2 2 2 2 2 √ 4 The splitting ﬁeld is E = Q( 2. i) √ Q( 3. Explicitly these are 1 1 1 1 1 1 1 1 √ + √ i. Q( 5. Irreducibility is a consequence of the polynomial version of the Eisenstein Test 1. so t is in fact a root of multiplicity p. Suppose that t ∈ k(T ) is a root of g(X). in fact we have Q(ζ5 ) = Q(cos(2π/5).
113
. i) and [E : Q] = 4. u3 = 3 2ζ3 . − + i. 3i) = Q( 6 2)( 3i) which has degree [E : Q] = 12.. 1. Then each automorphism √ α ∈ AutQ (Q( 3 2. − 2. i)
2 2
√ √ Q( 6 2. σα0 = (2 3).e. = σid = id.5. u2 = 3 2ζ3 . ζ3 )) permutes these roots. 6 2e2kπi/6 = 6 2ekπi/3 for k = 0. √ √ p2 (X) = X 6 − 2: The roots are the six complex 6th roots of 2. 3.
t ∈ K such that s ̸= t and K(u + sv) = K(u + tv). Notice that v ∈ K(u) and then f (uX + v) ∈ K(u). i)/Q: Here [Q( 2. 3. v) K(u + tv) K(u.16. Then if v ∈ C is a nonreal root of f (X) we have v ∈ Q(u). Consider the subﬁelds K(u + tv) K(u. We assume that K is inﬁnite since otherwise the result will be proved in Proposition 5. (b) Viewing Gal(E/K) as a subgroup of S3 . hence φQ(u) ̸= Q(u) and so Q(u)/Q is not normal.7. Chapter 4 4. Then there are only ﬁnitely many of these. so it is only necessary to show that the splitting ﬁeld E of p(X) over K is separable over K. so if we take u to be any cube root of c3 and u = −c2 /3c3 then f (uX + v) has the desired form.8. Here is the argument that K(u.1.
3. so there must be s. hence K(u. v) with t ∈ K.
114
. 3.4). so provided that we can ﬁnd a cube root of 1/c3 in K. so both its roots in K lie in E. v)/K is simple. Let E C be a splitting subﬁeld for f (X) over Q. by Theorem 4.81 we know that splitting ﬁelds are always normal. If E/K is a quadratic extension then for any u ∈ E − K we have 1 < [K(u) : K] 2. The example F2 (Z)/F2 (Z 2 ) is not separable since X 2 − Z 2 ∈ F2 (Z 2 )[X] is irreducible but not separable (see Qu.8 we know that 3 divides  Gal(E/K). √ √ √ Q( 4 3.7 for the rest of this question.u (X) must factor into linear factors over E. (a) Suppose that f (X) = c3 X 3 + c2 X 2 + c1 X + c4 with c3 ̸= 0. The induction is straightforward. (c) This is a tedious calculation! See Section 4. the result follows from Proposition 3. i) : Q] = 4 and the element 2 + i has degree 4 with minimal √ polynomial X 4 − 2X 2 + 9 which has roots ± 2 ± i.6. but the only subgroups of S3 with this property are S3 and A3 . By Theorem 3. so [K(u) : K] = 2 = [E : K] and therefore K(u) = E. i) : Q] = 4 and the element 3 + i has degree 4 with minimal √ polynomial X 4 − 4X 2 + 16 which has roots ± 3 ± i. √ √ √ Q( 3. Then (s − t)v = (u + sv) − (u + tv) ∈ K(u + tv). This shows that E is normal over K. C) = MonoQ (Q(u). i)/Q: Here [Q( 4 3.73 together with the fact that if L/K E/K is separable and v ∈ E is a root of p(X). 3. This implies that u = (u + tv) − tv ∈ K(u + tv). 4. i) : Q] = 8 and the element 4 3 + i has degree 8 with minimal √ √ polynomial X 8 + 4X 6 + 40X 2 + 4 which has roots ± 4 3 ± i and ± 4 3i ± i. hence v ∈ K(u + tv). / This means that there is a monomorphism φ ∈ MonoQ (Q(u).2. so f (X) does not split over Q(u) even though it has a root in this ﬁeld. If char K ̸= 2 then all quadratic polynomials over K are separable. we have f (uX + v) ∈ K. Then minpolyK. i)/Q: Here [Q( 3.√ √ √ Q( 2. Q) for which φ(u) = v. Then f (uX + v) = c3 u3 X 3 + (3c3 vu2 + c2 u2 )X 2 + (3c3 uv 2 + c1 u + 2c2 uv)X + (c3 v 3 + c4 + c1 v + c2 v 2 ). Since E is obtained by repeatedly adjoining roots of p(X). v) and so K(u. then L(v)/K is separable. v) = K(u + tv).
By the formula following Proposition 4. βα = {ι. If a/b is a rational root of f (X). βα. w)) is cyclic of order 3 whose generator is a 3cycle which cyclically permutes u. w)/Q) is a subgroup of S3 (viewed as the permutation group of {u. v. w)/Q) A3 ∼ Z/3. βα3 }. and these generate a dihedral
115
. βα = {ι. so we can take δ = 2 3i. w. − 4 3 ζ8 .4. w) = Q(u.26 implies that Gal(Q(u. w))/Q) = 3 = and Gal(Q(u. Using the choices of the proof. Since the discriminant is a square in Q.3. ⟨ ⟩ (b) The centre of D8 is α2 which has order 2. the discriminant of f (X) is ∆ = −27 − 4(−3)3 = 81 = 92 . so we may assume √ √ √ √ √ √ −1 243 243 4 4 u = 3 ζ8 = (1 + i). w) : Q]  3! = 6. This is an example of case (iii) of Kaplansky’s Theorem and we use the notation of the √ proof. Proposition 4. α . the splitting subﬁeld K(v. (a) This should be a familiar result. α3 }.5. If the distinct roots of f (X) in C are u. α2 . v. So  Gal(Q(u. v. The Galois group Gal(Q(u. v. but 1 is certainly not a root. b) = 1.25. v. βα}. ζ8 ) √ Q( 3. v. 2 2 √ √ where as usual ζ8 = e2πi/8 = (1 + i)/ 2. v. 4 3 ζ8 ) = Q( 4 3. 4. v. v. βα2 }. 4 3 ζ8 . Notice that there are also four nonnormal subgroups of order 2. w) : Q] and [Q(u. α2 . which easily implies that a. β = {ι. − 4 3 ζ8 of f (X) are given in permutation notation by σ = (1 4 3 2) and α = (1 2)(3 4). 4.4. we may assume that gcd(a. w. i)
j jjjj jjjj j jjjj jjj jj √ √ Q( 3i)
Q( 3) Q(i)
jjjj
jjjj
jj 2
jjjj2 2
jjjj
jjjj
Q
√ √ √ √ Then α is the restriction of complex conjugation to E. α2 . and there are three normal subgroups of order 4. α . α. This gives the diagram of subﬁelds of E √ √ −1 √ E = Q( 4 3 ζ8 . β( 3 ζ8 ) = β(− 3 ζ8 i) = − 3 ζ8 i. β}. Hence there are no rational roots and so no proper rational factors. The roots of X 2 + 3 are √ ± 3i. we have √ √ √ −1 √ √ 4 4 4 4 4 β( 3 ζ8 ) = − 3 ζ8 . hence also β(i) = −i. βα3 }. β. Now a3 − 3ab2 + b3 = 0. ⟨ 2⟩ ⟨ 3⟩ ⟨β⟩ = {ι. The discriminant here is δ 2 = −12. while β( 3i) = 3i and β( 3) = − 3. namely ⟨ 2 ⟩ ⟨ 2 ⟩ ⟨α⟩ = {ι. √ √ −1 √ √ −1 The eﬀects of σ and γ on the four roots 4 3 ζ8 . v = 3 ζ8 = (1 − i). w}). βα2 }. ⟨βα⟩ = {ι. βα = {ι. w) C satisﬁes 3  [Q(u. v. b = ±1. Hence we have uv = 3 and uvδ = 6i.
subgroup of S4 . Using the previous question (but beware that the notation there is inconsistent with that of the present situation!) we have the normal subgroups ⟨ 2⟩ ⟨ 2 ⟩ ⟨ 2 ⟩ σ , ⟨σ⟩ , σ ,α , σ , ασ , and these have ﬁxed ﬁelds √ 2 E ⟨σ ⟩ = Q( 3, i), √ ⟩ = Q( 3), √ ⟩ = Q( 3 i),
E ⟨σ⟩ = Q(i),
E ⟨σ
2 ,α
E ⟨σ
2 ,ασ
√ 4.6. Q(X 3 − 10)/Q: This is similar to Example 4.20, with splitting ﬁeld Q( 3 10, ζ3 ) and √ Gal(Q( 3 10, ζ3 )/Q) ∼ S3 . = √ √ √ √ √ √ √ Q( 2)(X 3 − 10)/Q( 2): The splitting ﬁeld is Q( 2, 3 10, ζ3 ), [Q( 2, 3 10) : Q( 2)] = 3 and √ √ √ √ 3 3 Q( 2, 10) Q( 2, 10, ζ3 ). √ √ √ Since ζ3 is not real, [Q( 2, 3 10, ζ3 ) : Q( 2)] = 6. The Galois group is isomorphic to S3 . √ √ √ Q( 3 i)(X 3 − 10)/Q( 3 i): Here Q( 3 i) = Q(ζ3 ), with [Q(ζ3 ) : Q] = 2. The splitting ﬁeld is √ √ √ Q( 3 10, ζ3 ) and [Q( 3 10, ζ3 ) : Q(ζ3 )] = 3, hence Gal(Q( 3 10, ζ3 )/Q(ζ3 )) ∼ Z/3 with generator σ = √ √ for which σ( 3 10) = 3 10 ζ3 . √ √ Q( 23 i)(X 3 − X − 1)/Q( 23 i): First note that X 3 − X − 1 ∈ Z[X] must be irreducible since its reduction modulo 2, X 3 + X + 1 ∈ F2 [X], has no root in F2 and hence has no linear factor (see Qu. 1.10). To proceed further we can use the ideas of Qu. 4.2 above (see also Section 4.7). √ The discriminant of the polynomial X 3 − X − 1 is ∆ = −23 and so δ = 23 i. Then if √ E = Q( 23 i)(X 3 − X − 1) is the splitting ﬁeld of X 3 − X − 1 over Q, Gal(E/Q) ∼ S3 and = √ ∼ A3 . Gal(E/Q( 23 i)) = √ √ K(X 3 − X − 1)/K for K = Q, Q( 5), Q( 5 i), Q(i): Continuing the preceding discussion, √ / notice that [E ∩ R : Q] = 3, so 5 ∈ E, hence √ √ Q( 5)(X 3 − X − 1) = Q(X 3 − X − 1)( 5) and √ √ [Q( 5)(X 3 − X − 1) : Q( 5)] = [Q(X 3 − X − 1) : Q] = 6, √ √ √ hence Gal(Q( 5)(X 3 − X − 1)/Q( 5)) ∼ S3 . Similarly, 5 i ∈ E and i ∈ E, hence / / = √ √ Gal(Q( 5 i)(X 3 − X − 1)/Q( 5 i)) ∼ S3 ∼ Gal(Q(i)(X 3 − X − 1)/Q(i)). = = 4.7. (a) Since char K ̸= 0, f ′ (X) = pX p−1 ̸= 0, so if u ∈ L is any root of f (X) then f ′ (u) = pup−1 ̸= 0. By Proposition 3.55, there are no multiple roots, hence p distinct roots. If u, v ∈ L are distinct roots, then (vu−1 )p = 1, so v = uξ for ξ ∈ K a pth root of 1 with ξ ̸= 1. (b) If there is a root u ∈ K, the Galois group Gal(L/K) acts in the following way. By The/ orem 4.8, there must be an element γ ∈ Gal(L/K) with γ(u) ̸= u. We can write γ(u) = uξγ r where ξγ ̸= 1 is a pth root of 1. Since γ(ξγ ) = ξγ , for r 1 we have γ r (u) = uξγ , which can only equal u if p  r. So u must have at least p conjugates which are all roots of f (X). Since deg f (X)p , every root of f (X) is conjugate to u, so f (X) must be irreducible over K. (c) Suppose that f (X) = g(X)h(X) with g(X) ∈ K[X] monic irreducible and 0 < d = deg g(X) < p. Let L/K with L a splitting ﬁeld for f (X) over K and let w ∈ L be a root of g(X). Arguing as in (a), we know that each root of g(X) has the form wξ where ξ is some
116
each of which is a normal extension of Q.
pth root of 1; moreover, L must contain p distinct pth roots of 1. Now the constant coeﬃcient of g(X) is g(0) = (−1)d ξ0 wd ∈ K where ξ0 is a pth root of 1. So
p g(0)p = (−1)dp ξ0 (wp )d = (−1)dp ad ,
from which it follows that ad is a pth power in K. As gcd(p, d) = 1, there are integers r, s such that rp + sd = 1, so we have a = (ar )p (ad )s = a pth power in K. Hence if f (X) is not irreducible in K[X] it has a root in K. 4.8. If u ∈ L is a root of f (X) in an extension L/K then by the Idiot’s Binomial Theorem 1.10 X p − a = X p + (−u)p = (X − u)p , so u is the only such root in L and f (X) splits over L. If (X − u)d ∈ K[X] for some d with 1 < d < p then ud ∈ K. Since gcd(d, p) = 1, there are integers r, s such that rd + sp = 1. Hence (ud )s (up )r = u, where the left hand side is in K. This shows that u ∈ K. Hence either f (X) has a root in K or it must be irreducible over K. Chapter 5 5.1. By Theorem 1.17, an integral domain D always admits a monomorphism into a ﬁeld j : D −→ F (e.g., F can be taken to be the ﬁeld of fractions of D), so any subgroup U D× becomes isomorphic to a subgroup jU F × , and if U is ﬁnite so is jU . Therefore jU and U are cyclic. 5.2. The only root of X 2 + 1 in F2 is the multiple root 1. 5.3. The ﬁeld Fpd [X]/(f (X)) is an extension of Fpd which has degree n, hence it is a ﬁnite ﬁeld with pdn elements, hence Proposition 5.6 implies that it is isomorphic to Fpdn . Since the extension Fpdn /Fpd is normal, Fpdn is a splitting ﬁeld for f (X) over Fpd . 5.4. (a) Here 41 is prime. Since 8  (41 − 1), there is a primitive 8th root of unity in F41 . 6 is a primitive root for F41 and 65 ≡ 27 (mod 4)1 has order 8. (b) Here 5 is prime 4  (5 − 1), so there is a primitive 4th root of unity in F× , but no primitive 5 8th root of unity. In fact, 2 and 3 have order 4, so these are primitive roots for F5 . Notice that in F5 [X], X 8 − 1 = (X 4 − 1)(X 4 + 1) = (X 4 − 1)(X 2 − 2)(X 2 − 3), where the polynomials X 2 − 2 and X 2 − 3 are irreducible. Therefore F25 is the splitting ﬁeld for X 8 − 1 over F5 and we have F25 ∼ F5 (u) = F5 (v), where u2 = 2 and v 2 = 3, so ±u and ±v are = primitive 8th roots of unity. To ﬁnd an element of order 24 in F× , we ﬁrst ﬁnd one of order 3. 25 Consider the polynomial X 2 + X + 1 ∈ F5 [X]; in F5 , this has roots which have order 3. These roots are given by (−1 ± w)/2, where w2 = (1 − 4) = −3 = 2, hence they are (−1 ± u) = −3 ± 3u. 2 Now the elements ±(2 ± 2u)u = ±(±4 + 2u) = ±4 ± 2u all have order 8 × 3 = 24. (c) Here 11 is prime and 8  (121 − 1) = 120, so F121 is the splitting ﬁeld of X 8 − 1 over F11 . The polynomial X 2 + 1 is irreducible over F11 so F121 = F11 (z) where z 2 = −1. Since 120 = 8 × 3 × 5, it is suﬃcient to ﬁnd elements of order 8, 3 and 5 whose product will have order 120.
117
Suppose that a + bz ∈ F121 with a, b ∈ F11 . If this element has order 8, then (a + bz)2 = ±z. So let us solve (a2 − b2 ) + 2abz = z. Then 2ab = 1 and b2 = a2 , hence b = ±a. Now we have 2a2 = ±1 and so a2 = ±1/2 = ±6. Now 6 is not a square in F11 but 72 ≡ −6 ≡ 42 (mod 11),
so we have a = 4, b = ±4 and a = 7, b = ±7. Therefore the elements of order 8 in F× are 121 4 ± 4z and 7 ± 7z. By the same approach as in (b), the elements of order 3 in F121 are (−1 ± 5z)/2 = 5 ± 8z. 2 is a primitive root for F11 so 4 = 22 has order 5. Combining these we obtain the following primitive roots for F121 : 7 ± z, 10 ± 4z. (d) In F2 [X] we have X 8 − 1 = (X − 1)8 , whose only root in F2 is 1. So the splitting ﬁeld is F2 . 5.5. Notice that Fp (w) is a splitting ﬁeld of the separable polynomial X p −1 − 1 over Fp , so if w ∈ F× then Fp (w) Fpℓ . Since Fp (w) = Fpd we have d ℓ; we also have degFp w = d. pℓ The number of conjugates of w is d, hence each primitive root of Fpd has d conjugates and the total number of these is the number of generators of the cyclic group F× ∼ Z/(pd − 1), i.e., pd =
d
φ(pd − 1). Hence d  φ(pd − 1). This can also be interpreted in terms of the evident action of Gal(Fpd /Fp ) ∼ Z/d on the set of all primitive roots of Fpd ; each orbit has exactly d elements, = so the number of orbits is φ(pd − 1)/d which is an integer.
′ 5.6. (a) First note that ga (X) = −1, so ga (X) is separable, hence E/K is separable. If u ∈ E is a root of ga (X), then for t ∈ Fpd ,
ga (u + t) = (u + t)p − (u + t) − a = (up − u − a) + (tp − t) = (up − u − a) = 0, hence u + t is also a root of ga (X). This means that E = K(u) since all the other roots of ga (X) lie in K(u). As ga (X) is irreducible over K, [E : K] = pd =  Gal(E/K) and so the following pd automorphisms are the elements of Gal(E/K): σt : E −→ E; σt (u) = u + t (t ∈ Fpd ).
d
d
d
d
It is easy to check that for s, t ∈ Fpd , σs ◦σt = σs+t . Hence there is an isomorphism Gal(E/K) ∼ = Fpd with σt corresponding to t ∈ Fpd . (b) If ga (X) is irreducible over K then it cannot have a root in K since its degree is greater than 1. Conversely, suppose that ga (X) has no root in K. Then if u ∈ E is any root of ga (X) in a splitting ﬁeld over K, the other roots are the p elements u + t ∈ E (t ∈ Fp ). If u + t0 ̸= u is a conjugate of u with 0 ̸= t0 ∈ Fp , there must be an element τt0 ∈ Gal(E/K) for which τt0 (u) = u + t0 . Then ⟨τt0 ⟩ must be isomorphic to a nontrivial subgroup of Fp , but this must be Fp since this group is simple. Hence, u must have p conjugates and so ga (X) is irreducible over K. (c) If K is a ﬁnite ﬁeld and d > 1 then if ga (X) were irreducible over K, then by (a), E would be ﬁnite and Gal(E/K) ∼ Fpd . But Fpd is not cyclic, yet we know from Proposition 5.23 that = Gal(Fpd /Fp ) ∼ Z/d is cyclic. =
118
. 12 + a2 + b2 = 0. p (b) If u ∈ Σq . Thus for every t ∈ Fq . we may therefore take λ2d (u) = 1 for all u ∈ F× . In this situation. which is even. then either u = 0 or u ̸= 0 and u = (±v)2 for some v ∈ F× . Then Fpd  = p pd normal subgroup (F× )2 = {u2 : u ∈ F× } pd pd F× /(F× )2 ∼ {±1}. G ∼ Z/p.12. Suppose that the result holds whenever = k with k < n. there are u. If ζ ∈ K is a primitive nth root of unity. there is a normal subgroup N ▹ G containing Z and satisfying N/Z = M ⊆ G/Z. Thus we have q Σq  = 1 + Then t − Σq  = Σq  = (c) Since Σq ∪ (t − Σq ) ⊆ Fq . while (−ζ)2n = (−1)2n ζ 2n = 1. which is odd.1. (a) By Proposition 5. (q + 1) . v ∈ Fq (possibly 0) for which u2 = t − v 2 . F× is a cyclic group.5. i. Clearly we have ker λq = (F× )2 .7. = pd pd We may use this group isomorphism to deﬁne λq .2. recall that by Cauchy’s Lemma. we may write −1 = a2 + b2 for some a. Hence −ζ ∈ K is a primitive 2nth root of unity. By the inductive hypothesis. (q + 1) − Σq ∩ (t − Σq ) Σq ∪ (t − Σq ) = Σq  + t − Σq  − Σq ∩ (t − Σq ). hence
Chapter 6 6. So now d 2 2d × d − 1. then (−ζ)n = (−1)n ζ n = −1.
119
. If p = 2 then F×  = 2d − 1. we have q This implies that q and so Σq ∩ (t − Σq ) 1. 6. Now when n = 1.e. The set of squares in F× is the suppose that p is odd. Now if G = pn . This establishes the inductive step and hence the desired result. q 2d so every element of F× is a square. the Legendre symbol of u from Number Theory. pd λq is surjective if and only if p ( odd. which is abelian. Hence G/Z has order G/Z = pk with k < n. Clearly N  = Z M  = pn−1 . λp (u) = . 2 2 F× pd and it is easily seen that its quotient group has order 2. By one of the Isomorphism Theorems. (d) By (c). is ) u Remark: when d = 1. whence t = u2 + v 2 . −t ̸= t (since otherwise 2t = 0 and so t = 0). 2 (q − 1) (q + 1) = . there is a normal subgroup M ▹ G/Z with M  = pk−1 . b ∈ Fq . for any nonzero t ∈ K. the centre Z of G G = p is nontrivial.
For any n 1. s = 1 and k = 0 (hence n = 5). As φ(5) = 4. 3 (hence n = 1. the residue class modulo 24.6. The eﬀect r of these elements on Q(ζ24 ) is given by r · ζ24 . Notice that 23 acts like complex conjugation. = 6. s = 1 and k = 0. √ √ 1 3 Q( 3 i): This contains the six 6th roots of unity ±1. 19. 1. We also have
−1 ζn − ζn = 2 sin(2π/n) i ∈ Q(ζn ). we have [Q(ζn ) : Q] = φ(2k) = φ(2)φ(k) = φ(k). p1 < p2 < · · · < ps . We have φ(20) = φ(4)φ(5) = 2 × 4 = 8 and the elements of (Z/20)× are the residue classes modulo 20 of the numbers 1. 13. where each pj is an odd prime. (a) We have φ(24) = φ(8)φ(3) = 4 × 2 = 8. 1. • p1 = 3. 19. so we might as well assume that n is even from now on. From this we see that the only roots of unity in Q(i) are ±1. for instance 7 and 13. r.5. 5. hence these all have order 2 except 1 which has order 1. • s = 0 and k = 0. 11. 2 (hence n = 3. The elements of Z/24 which are invertible are the residue classes modulo 24 of the numbers 1. 9. i) Q(i) which is impossible since [Q( 3. Then
r φ(n) = φ(2k )φ(pr1 ) · · · φ(prs ) = φ(2k )(p1 − 1)p11 −1 · · · (ps − 1)prs −1 . 2. 3. 17. 6.
(a) If 4 n then writing n = 2k with k odd. rj s 1 k 0.3. 4. 12). This time there are elements of order 4. satisﬁes r2 = 1. Q(i): Here degree [Q(i) : Q] = 2 and clearly the four 4th roots of unity ±1. 17. s s 1
1 and
If s > 0 then φ(n)  4 happens precisely when r1 = · · · = rs = 1 and one of the following possibilities occurs: • p1 = 5. (b) This is similar to (a). let ζn = e2πi/n = cos(2π/n) + sin(2π/n) i. while [Q(ζ2n ) : Q] = φ(4k) = φ(4)φ(k) = 2φ(k). so in particular. ±i lie in this ﬁeld. 8). Since (Z/24)× is abelian. √ Q( 2 i): This ﬁeld contains only the square roots of unity ±1. 23. For each of these numbers r. 2 2 √ Q( 5 i): This ﬁeld contains only the square roots of unity ±1. 7. The eﬀect on Q(cos(π/12)) is given by r · cos(π/12) = cos(πr/12). ±i. 13. Write n = 2k pr1 · · · prs . i) : Q] = 4.
120
. it is isomorphic to Z/2 × Z/2 × Z/2.4. If it contained a 3rd root of unity then it √ √ √ would contain 3 and so Q( 3. 7. 6. it has no 5th roots of unity except 1. ± ± i. −r · cos(π/12) = cos(−πr/12) = cos(πr/12) = r · cos(π/12). then Q(ζn ) = Q(−ζn ) where −ζn is a primitive 2nth root of unity. Notice that if n is odd. Then we have (Z/20)× ∼ Z/2 × Z/4. 11.
we have [Q(cos(π/n)) : Q] = 2[Q(cos(2π/n)) : Q].
ℓ If 4  n. [Q(cos(π/ℓ))(sin(π/2ℓ)) : Q(cos(π/ℓ))] = 2 and we must have Q(cos(2π/n)) = Q(cos(π/2ℓ)) = Q(sin(π/2ℓ)) with Gal(Q(sin(π/2ℓ))/Q) = Gal(Q(cos(π/2ℓ))/Q) = (Z/4ℓ)× /{1. −1} (b) We have √ 2− 3 1 − cos(π/6) = . it is easy to Consider the automorphism σ ∈ Gal(Q(ζn )/Q)) for which σ(ζn ) = ζn n see that σ has order 2. whence −1 ζn − ζn ∈ Q(ζn ). So k we see that sin(2π/n) ∈ Q(ζn ) in this situation. sin (π/12) = 2 4
2 121 −1 −ζn + ζn 2(−ζn )ℓ
sin(π/2ℓ) if ℓ is odd.k Hence. σ(cos(π/ℓ)) = cos(π/ℓ). 2i
sin(2π/n) ∈ Q(ζ2n ) ∩ R = Q(cos(π/n)). if ℓ is even. /
sin(2π/n) = and by Theorem 6. −1}.
2ℓ+1 = −ζ . hence Q(cos(π/n)) = Q(cos(2π/n))(sin(2π/n)) and [Q(cos(2π/n))(sin(2π/n)) : Q(cos(2π/n))] = 2. Notice that since i = ζ2n . σ(sin(2π/n)) = σ(sin(π/2ℓ)) = From this we ﬁnd that when ℓ is odd.
.3. = − sin(π/2ℓ) if ℓ is even. Also. with minpolyQ(cos(2π/n)). Thus we have [Q(sin(π/2ℓ)) : Q] = 2φ(ℓ) and Gal(Q(sin(π/2ℓ))/Q) = Gal(Q(cos(π/2ℓ))/Q) = (Z/4ℓ)× /{1. Then
σ(cos(2π/n)) = σ(cos(π/2ℓ)) = − cos(π/2ℓ). we can write n = 4ℓ. Then i = ζn . Q(ζn ) cannot contain ζ2n and by another simple argument it cannot contain i = ζ2n . Q(cos(2π/n)) = Q(cos(π/2ℓ)) = Q(cos(π/ℓ))(sin(π/2ℓ)) = Q(sin(π/2ℓ)).sin(2π/n) (X) = X 2 + cos2 (2π/n) − 1. so i ∈ Q(ζn ). Similarly.
−2 2 ζ2n − ζ2n ∈ Q(ζ2n ). i
sin(π/2ℓ) = sin(2π/n) = Clearly
sin(π/2ℓ) ∈ Q(ζn ) ∩ R = Q(cos(2π/n)). since cos(π/ℓ) = 1 − 2 sin2 (π/2ℓ) ∈ Q(sin(π/2ℓ)).
(a) We have  Gal(Q(ζ5 )/Q) = [Q(ζ5 ) : Q] = deg Φ5 (X) = φ(5) = 4. 5 · 3 = 3. 4 √ √ − 6− 2 . and (Z/5)× is cyclic generated by the residue class 2. 1 · 3 = 3. = =
and Here the eﬀect of the coset of the residue class of r ∈ (Z/4ℓ)× is given by r · sin(π/12) = Explicitly we have √ √ 6− 2 1 · sin(π/12) = −1 · sin(π/12) = sin(π/12) = . 4 2 · ζ5 = ζ5 .
Hence 4 cos2 (2π/5) + 2 cos(2π/5) − 1 = 0. √ √ √ √ √ √ √ √ 7 · 2 = 2. −2 −1 2 (ζ5 + ζ5 ) + (ζ5 + ζ5 ) + 1 = 0
and therefore
−1 −1 (ζ5 + ζ5 )2 + (ζ5 + ζ5 ) − 1 = 0. 2 4
122
. 4
√ √ √ √ Q(sin(π/12)) = Q( 6 − 2) = Q( 2. 3 r r ζ24 − ζ24 = sin(rπ/12) i1−r . The action is given by
2 2 · ζ5 = ζ5 . 2 3 2 · ζ5 = ζ5 .6. 6. 5 · 3 = − 3. −1} ∼ Z/2 × Z/2. 5 · 2 = − 2. 11 · 2 = − 2. so this is this polynomial is irreducible over Q. The quadratic polynomial 4X 2 + 2X − 1 ∈ Z[X] has discriminant 20 which is not a square in Q.cos(2π/5) (X) = X 2 + X − . 11 · 3 = − 3. therefore 1 1 minpolyQ. ir
2 · ζ5 = ζ5 . so since ζ5 = ζ5 and ζ5 = ζ5 .
4
−1 −2 −1 3 4 (b) We have ζ5 + ζ5 = 2 cos(2π/5) and Φ5 (ζ5 ) = 0. 3). Gal(Q(sin(π/12))/Q)) ∼ (Z/4ℓ)× /{1. 4 √ √ In terms of the generators 2 and 3 these act by √ √ √ √ √ √ √ √ 1 · 2 = 2.and so sin(π/12) = Then
√
2− 2
√ 3
√ √ 6− 2 = . 7 · sin(π/12) = −7 · sin(π/12) = − sin(7π/12) = 4 √ √ − 6+ 2 11 · sin(π/12) = −11 · sin(π/12) = − sin(11π/12) = . 4 √ √ 6+ 2 5 · sin(π/12) = −5 · sin(π/12) = sin(5π/12) = .
(c) Taking square roots we ﬁnd that ±√p ξ= ±√p i As ξ ∈ Q(ζp ). (b) Since −1 if p ≡ 1 (mod 4). 8 ∼ Z/4 and has 3 subgroups {1} {1. 4} Gal(Q(ζ5 )/Q). Q(ζ5 )
2
√ Q(ζ5 )⟨4⟩ = Q(cos(2π/5)) = Q( 5)
2
Q 6.√ √ −1 ± 5 −1 + 5 The roots of this are . (a) We have
(p−1)/2
ξ2 =
∏
(p−1)/2 r −r −r r (ζp − ζp )(ζp − ζp )
r −r (ζp − ζp )2 = (−1)(p−1)/2
∏
r=1
r=1 (p−1)/2
= (−1)(p−1)/2 = (−1)(p−1)/2
∏
−2r 2r (1 − ζp )(1 − ζp )
r=1 p−1 ∏ r=1 (p−1) 2r (1 − ζp )
= (−1)(p−1)/2
∏
s (1 − ζp )
s=1
since each congruence 2x ≡ t (mod p) has exactly one solution modulo p for each t. 4 √ √ 5+ 5 1+5−2 5 2 2 = . sin (2π/5) = 1 − cos (2π/5) = 1 − 16 8 √ √ 5+ 5 hence sin(2π/5) = .
the result follows. and
p−1 ∏ s=1 s (1 − ζp ) = Φp (1) = p. (p−1)/2 (−1) = 1 if p ≡ 3 (mod 4). As sin(2π/5) > 0. We 4 √ 4 −1 − 5 also have cos(4π/5) = .7. if p ≡ 3 (mod 4).
123
. we see that
if p ≡ 1 (mod 4).
√ √ p ∈ Q(ζp ) if p ≡ 1 (mod 4) and p i ∈ Q(ζp ) if p ≡ 3 (mod 4). giving the following (c) Gal(Q(ζ5 )/Q) = tower of subﬁelds. As cos(2π/5) > 0 we must have cos(2π/5) = .
Also recall that every permutation ρ ∈ Sn is a product of 2cycles. so there is an element t ∈ E for which TrE/K t = t + σ(t) + · · · + σ n−1 (t) ̸= 0. s[m] = Xir . So we obtain v= 1 TrE/K t u−σ ( ) u . so it suﬃces to show that every 2cycle (a b) ∈ Sn is a product of 2cycles of the form (r + 1 r + 2). Write ∑ ∑ e[m] = Xi1 · · · Xir .10. Then u = vσ(t) + (v + σ(v))σ 2 (t) + · · · + (v + σ(v)σ 2 (t) + · · · + σ n−2 (v))σ n−1 (t) satisﬁes ( ) ( ) u − σ(u) = v σ(t) + σ 2 (t) + · · · + σ n−1 (t) − σ(v) + · · · + σ n−1 (v) t ( ) ( ) = v t + σ(t) + σ 2 (t) + · · · + σ n−1 (t) − v + σ(v) + · · · + σ n−1 (v) t = (TrE/K t)v − (TrE/K v)t = (TrE/K t)v. Recall the wellknown formula σ(i1 · · · ir )σ −1 = (σ(i1 ) · · · σ(ir )). (a) For each u ∈ E. σ(T (u)) = σ(u + σ(u) + σ 2 (u) + · · · + σ n−1 (u)) = σ(u) + σ 2 (u) + · · · + σ n (u) = σ(u) + σ 2 (u) + · · · + σ n−1 (u) + u = T (u). while (1 2 · · · n)n−1 (1 2)((1 2 · · · n)n−1 )−1 = (1 2 · · · n)−1 (1 2)((1 2 · · · n)−1 )−1 = (n 1) = (1 n). we also have (a b) = (b − 1 b) · · · (a + 2 a + 3)(a + 1 a + 2)(a a + 1)(a + 1 a + 2)(a + 2 a + 3) · · · (b − 1 b).8. the linear combination of characters id +σ + · · · + σ n−1 must be linearly independent. Assuming that a < b. 6.6. Hence H = Sn . It is straightforward to verify that the resulting function TrE/K : E −→ K is Klinear. By Artin’s Theorem 6. Then for 1 r n − 2 we have
(1 2 · · · n)r (1 2)(1 2 · · · n)n−r = (1 2 · · · n)r (1 2)((1 2 · · · n)r )−1 = (r + 1 r + 2). and this is in H. (a) This can be proved by induction on n. r r
i1 <i2 <···<ir m 1 i m
Then we easily ﬁnd that
[m] er = e[m−1] + er−1 Xm .
1 TrE/K t
6. so T (u) is ﬁxed by σ and all its powers. hence by Gal(E/K).15. Therefore T (u) is in E Gal(E/K) = K.9. This means that every such 2cycle (r + 1 r + 2) is in H. (b) Let v ∈ E and suppose that TrE/K (v) = 0. r 124 [m−1] [m] r sr = s[m−1] + Xm . r
.
which demonstrates the inductive step. Now suppose that [n+1] [n] k the result is true for some n 1. The desired result is that for all n
[n] [n] [n] [n] [1] [n] [n] [n]
1 and k
1. (b)(i) We have h1 = e1 . 1 1 (ii) This can be done by induction on n in a similar way to part (a). Then sk = sk + Xn+1 . while [1]
[n]
e1
[n+1] [n+1] sk−1 [n]
− e2
[n+1] [n+1] sk−2 [n]
+ · · · + (−1)k−1 ek−1 s1
[n] [n] [n] [n]
[n+1] [n+1]
+ (−1)k kek
[n]
[n+1]
=
k−1 k−2 (e1 + Xn+1 )(sk−1 + Xn+1 ) − (e2 + e1 Xn+1 )(sk−2 + Xn+1 ) + · · ·
+ (−1)k−1 (ek−1 + ek−2 Xn+1 )(s1 + Xn+1 ) + (−1)k k(ek + ek−1 Xn+1 )
k−1 k−2 = sk + (e1 Xn+1 − e2 Xn+1 + · · · + (−1)k−1 ek−1 Xn+1 ) [n] [n] [n] [n] [n] [n] [n] [n] [n] [n]
[n]
[n]
[n]
+ (sk−1 − e1 sk−2 + · · · + (−1)k−1 ek−2 s1 + (−1)k kek−1 )Xn+1
k−1 k 2 + (Xn+1 − e1 Xn+1 + · · · + (−1)k−1 ek−2 Xn+1 ) k = sk + Xn+1 = sk [n] [n+1] [n] [n]
.
125
.
r When n = 1 we have sr = X1 and e1 = X1 from which the result follows. h2 = e2 − e2 and h3 = e3 − 2e1 e2 + e3 .Notice also that er
[m]
= 0 whenever r > m.
sk = e1 sk−1 − e2 sk−2 + · · · + (−1)k−1 ek−1 s1 + (−1)k kek .