Lecture Notes on Fields

Hong-Jian Lai

March 2003

1. Field Extensions

(1.1) If K is a subfield of F , then F is an extension field of K; and we denote

this fact by K ≤ F .

(1.2) Examples: Z2 ≤ Z2 [x]/(x2 + x + 1), Q ≤ R ≤ C, ...

(1.3) Let K ≤ F be fields. An element α ∈ F is algebraic over K if ∃f (x) ∈ K[x]

such that f (α) = 0. If α is not algebraic over K, then α is transcendental over K.
In particular, a complex number which is algebraic over Q is an algebraic number;
a complex number which is not algebraic over Q is a transcendental number.

F is an algebraic extension of K if every element of F is algebraic over K; and F

is a transcendental extension of K if at least one element of F is transcendental
over K.
Example: There exist transcendental elements in R over Q. In fact, R is uncountable
whereas the set of all algebraic numbers over a countable field is countable.

(1.4) Must every f (x) ∈ K[x] have a root?

Kronecher’s Theorem: YES, in some extension field of K.

Proof (1) We may assume that f (x) is irreducible.

(2) Then K[x]/(f (x)) is a field in which x + (f (x)) is an element.
(3) Show that α = x + (f (x)) is a root of f (x).

1
(1.5) Let K ≤ F be fields, and let X ⊆ F be a subset. Let K(X) denote the
intersection of all subfields of F that contains K ∪ X. If X = {α}, then write K(α)
for K({α}). If α ∈ F − K, then K(α) is a simple extension of K.

(1.6) If α ∈ F − K is algebraic over K, then each holds:

(i) there is an irreducible polynomial p(x) ∈ K[x] such that p(x) = 0; and
(ii) if q(x) ∈ K[x] is another irreducible polynomial with q(x) = 0, then p(x) =
cq(x) for some c ∈ K. (Therefore, there is only one monic irreducible polynomial
irr(α, K)(x) ∈ K[x] such that irr(α, K)(α) = 0 for each algebraic α over K. The
degree of irr(α, K) is the degree of α over K, denoted deg(α, K).

(1.7) Structure of simple extensions: Let K ≤ F be fields and let α ∈ F − K.

Then each holds:
(i) If α is transcendental over K, then K(α) is isomorphic to the field of quotients of
the ID K[x].
(ii) If α is algebraic over K, then K(α) is isomorphic to the factor K[x]/(irr(α, K)).
Proof Let φα : K[x] → K(α) by φ(f (x)) = f (α). Use 1st isomorphism theorem for
rings.

(1.8) Let K ≤ F be fields. Then F can be viewed as a vector space over K. The
dimension of F as a vector space over K is denoted by [F : K] (regardless whether
F is finitely dimensional or not).
(i) Let K ≤ F ≤ E be fields, then [E : K] = [E : F ][F : K].
(ii) [E : K] < ∞ if and only if both [E : F ] < ∞ and [F : K] < ∞.
Proof:

(1.9) Let α ∈ F be algebraic over K with deg(α, K) = n. Then each holds:

(i) K(α) is a vector space with {1, α, · · · , αn−1 } as a basis. (In other words, dim
K K(α) = n = deg(irr(α, K)).)
(ii) Every element β ∈ K(α) is algebraic over K with deg(β, K) ≤ deg(α, K).

√ √
Example: Q( 2), Q( 3 3).

2
(1.10) Let K ≤ F, L ≤ J be fields, and σ : K 7→ L be a field isomorphism, and
u ∈ F − K and v ∈ J − L. Suppose that
(i) u is transcendental over K and v is transcendental over L; or
(ii) u is algebraic over K and v is algebraic over L such that σ(irr(u, K)) = irr(v, L).
Then there exists a filed isomorphism K(u) ∼ = L(v) which extends σ.
Proof: σ extends to a ring isomorphism K[x] ∼
= L[x]. If (i) holds, then (1.10) follows
from (1.7)(i). If (ii) holds, then (1.10) follows from (1.7)(ii).

(1.11) Let K ≤ F and K ≤ E be fields, u ∈ E and v ∈ F are algebraic over

K. Then u and v are roots of the same irreducible polynomial p ∈ K[x] if and only if
there exists a filed isomorphism φ : K(u) ∼
= K(v) such that φ|K is the identity map
of K.
Proof: (ONLY IF) σ = 1K in (1.10).
(IF) 1K (p) = p (for the coefficients of p).

(1.12) Let K be a field and f ∈ K[x]. Then there exists a field F = K(u) such
that
(i) u ∈ F is a root of f ;
(ii) [F : K] ≤ deg(f ).
(iii) If f is irreducible in K[x], then K[u] is unique up to an isomorphism which is
the identity of K.
Proof: Replace f by an irreducible factor if needed, we may assume that f is irre-
ducible. Then (f ) is a maximal ideal in K[x], and so K[x]/(f ) ∼
= F := K[u] for some
root u of f , (see (1.7)(ii)) with [F : K] = deg(f ). If there is another field K(v) such
that v is a root of f , then by (1.11), there is a filed isomorphism K[u] ∼ = K[v] which,
when restrict to K, is the identity of K.

(1.13) Let K ≤ F be field. If [F : K] = n < ∞, then F is an algebraic exten-

sion of K. Moreover, there exists a finite set X ⊆ F such that F = K(X) (in this
case, we say that F is finitely generated over K; and when [F : K] = n < ∞, we
say that F is a finite extension of K).

3
Proof: Use a basis of F , as a vector space over K. If [F : K] = n, then 1, α, · · · , αn
must be linearly dependent.

Example: If K ≤ F and F ≤ E are finite extensions, then [E : K] = [E : F ][F :

K] < ∞, and so F ≤ K is also a finite extension. In particular,
√ √ √ √ √
Q( 2 + 3) = (Q( 2))( 3) ≥ Q( 2) ≥ Q.

(1.14) Let K ≤ F be fields such that F = K(X).

(i) If every element in X is algebraic over K, then F is an algebraic extension over
K.
(ii) If, in addition, |X| < ∞, then [F : K] < ∞.
Proof: Apply (1.8).

(1.15) Let K ≤ F ≤ E be fields. If F is algebraic over K, and if E is algebraic

over F , then E is algebraic over K.
Proof: Use definition and (1.8) (or (1.14)).

2. Algebraically Closed Fields and Algebraic Closure

(2.1) Let F ≤ E. Then

F̄E = {α ∈ E|α is algebraic over F }

is a subfield of E, called the algebraic closure of F in E.

Proof ∀α, β ∈ F̄E , show that α+β, α−β, αβ, α/β ∈ F̄E , by working in F (α, β) ⊂ F̄E .

(2.2) A field F if algebraically closed if every nonconstant polynomial in F [x]

has a zero in F . An algebraic closure of F is an algebraic extension F̄ that is
algebraically closed.

(2.3) A field F is algebraically closed if and only if every f (x) ∈ F [x] − F can
be factored into a product of linear factors.

4
Proof Definition and Induction on the degree.

(2.4) If F is algebraically closed, and if F ≤ E is an algebraic extension, then F = E.

(2.5) Main Theorem Every field has an algebraic closure.

(2.5) Review the concepts: POSETs, chains, upper bounds of a subposet, maxi-
mal element. Linear ordered sets, Well-ordered sets (every nonempty set has a least
element).

(2.6) Zorn’s Lemma: If (S, ≤) is a poset in which every chain has an upper bound,
then S has at least one maximal element.

Non-Example: (Z, ≤) does not have a maximal element, but it is well ordered.

Remark: Zorn’s Lemma is equivalent to the following:

(i) Zermelo’s Axiom of Choice Let M be a set and let S the the power set of M ,
then there is a choice function f : S → M such that f (M 0 ) ∈ M, ∀M 0 ∈ S.
(ii) Well-Ordering Theorem Any set can be well-ordered.
(iii) Hausdorff Maximality Principle Every non-empty poset contains a maximal
chain.
(iv) Turkey’s Lemma Every non void family of finite character has a maximal mem-
ber. (A family F of sets is said to have finite character if for each set A, A ∈ F if
and only if every finite subset (including the empty set) of A is in F.)

(2.7) Proof of (2.5) Let F be a field.

Discussion: Let S denote the collection of all algebraic extensions of F , partially
ordered by ≤ (field extension). Then verify the condition of Zorn’s Lemma.
However, this does not work well by Russell’s paradox: since the collection of all
algebraic extension of F may not be a set! Thus S must be carefully selected.
Russell’s Paradox: Consider the class M = {X|X is a set and X 6∈ X}. If M is a
set, then either M ∈ M , whence M 6∈ M , or M 6∈ M , whence M ∈ M .

(2.8) For any set X, there exists a set Y with |Y | > |X|.

5
Proof: Take Y to be the power set of X (that is, Y is the set of all subsets of X).

Proof of (2.6). Let F be a field.

(i) Form the Ω.

Define
A = {ωf,i : f ∈ F [x], i = 0, 1, · · · , degree of f }

to be a set that has an element for every possible zero of any f (x) ∈ F [x]. Define Ω
to be a set with |Ω| > |A|. We may assume that F ⊂ Ω (otherwise we replace Ω by
Ω ∪ F ).

(ii) Form the POSET S.

Consider all the fields that are algebraic extensions of F . We can view them all as
subsets of Ω (If α is a root of f (x) ∈ F [x] with irr(α, F ) = a0 + a1 x + · · · + an xn ,
then rename α be ω and renaming elements in F (α) by the corresponding element in
F (ω), we can view F (α) = F (ω) ⊆ Ω. Since each degree n polynomial f (x) ∈ F [x]
can have at most n root, and by the definition of A, this can be done).

Let S be the set of all algebraic extension fields Ej of F such that Ej ⊆ Ω:

S = {Ej : j ∈ J, F ≤ Ej ⊆ Ω}.

S becomes a poset with the set containment relation: Ei ≤ Ej ⇐⇒ Ei ⊆ Ej .

(iii) Verify Zorn’s Lemma.

Let T = {Eik } be a chain in S, and let W = ∪k Eik . We shall show that W is also an
algebraic extension field of F .
Let ω1 , ω2 ∈ S with ω2 6= 0. Then ∃Ei1 , Ei2 such that ω1 ∈ Ei1 and ω2 ∈ Ei2 .
Since T is a chain, we may assume that Ei1 ⊆ Ei2 . and so ω1 ± ω2 , ω1 ω2 and ω1 /ω2 ∈
Ei2 ⊆ W . Similarly, We can routinely verify that W satisfies the axioms of a field.
Thus W is an extension field of F .
For each ω ∈ W , ∃Ei ∈ S such that ω ∈ Ei . Since Ei is an algebraic over F , ω is
algebraic over F . Hence W ∈ S, and so W is an upper bound for T .
By Zorn’s Lemma, there must be a maximal element F in S.

6
(iv) Show that F is an algebraic closure of F .
Let f (x) ∈ F [x], where f (x) 6∈ F . Suppose that f (x) has no root in F . Since
|Ω| > |A|, we can pick ω 0 ∈ Ω such that f (ω 0 ) = 0 and F (ω 0 ) ⊆ Ω.
(Show that F (ω 0 ) is an algebraic extension of F .) Let β ∈ F (ω 0 ). Then β is
algebraic over F and so for some α0 , α1 , · · · , αn ∈ F such that

α0 + α1 β + · · · + αn β n = 0.

Therefore, αi is algebraic over F , and β is algebraic over F (α0 , α1 , · · · , αn ). It follows

that F (α0 , α1 , · · · , αn , β) is a finite (and so algebraic) extension over F . Therefore,
F (β) is algebraic over F and F ⊂ F (β), contrary to the maximality of F . Therefore,
F must be algebraically closed.

Finite Fields
(F1) If F is a finite field with char(F ) = p, then there is an integer n ≥ 1 such that
|F | = pn . (Proof: Use vector space.)

(F2) If p is a prime and n ≥ 1 is an integer, then Zp has an extension F such

that |F | = pn . (Notation: F = GF (pn ).)
n
Proof: Step 1: Show that every root of xp − x ∈ Zp [x] is simple. Let α be a zero of
n −1 n −1
xp − 1. Then αp = 1. Consider
n −1 n −1 n −1
xp − 1 = xp − αp
n −1 n −2 n −2 n −3 n −2 n −1
= xp − αxp + αxp − α 2 xp + · · · + αp x − αp
n −2 n −3 n −4 n −3 n −2
= (x − α)(xp + αxp + α 2 xp + · · · + αp x + αp )
= (x − α)g(x)

Note that
n −2
g(α) = (pn − 1)αp = −α−1 6= 0.
n
Hence the factor of x − α occurring in xp − x only once.

Step 2: Let K be a subset of Zp such that

n
K = { all zeros of xp − x}.

7
Note that Zp ⊂ K. We want to show that K is a subfield of Zp . Note if α, β(6= 0) ∈ K,
then
n n n
(α − β)p = αp − β p = α − β
n n n
(αβ −1 )p = αp (β p )−1 = αβ −1

(F3) Zp = ∪n≥1 GF (pn ).

Proof: From (F2), one can see that

GF (pm ) ⊆ GF (pn ) ⇐⇒ m|n.

Let E = ∪n≥1 GF (pn ). Define addition and multiplication on E as follows: ∀a, b ∈ E,

we may assume that a ∈ GF (pn1 ) and b ∈ GF (pn2 ). Thus a, b ∈ GF (pn1 n2 ) and so
we define a + b and ab as they are in GF (pn1 n2 ).

Pm
To see that E is algebraically closed, pick f (x) = i=0 ai xi ∈ E[x]. Then there
ni ’s such that ai ∈ GF (pni ), and so all a0 , · · · , am ∈ GF (ps ), where s = Πm
i=0 ni . As
[GF (ps ) : Zp ] = t < ∞, all zeros of f (x) are algebraic over Zp and are in some GF (pt ),
and so are in E.

By (F2), each GF (pn ) is algebraic over Zp and so E is algebraic over Zp .

8
3. Automorphisms of Fields

(3.1) Let E, F be extension fields of a field K. A field homomorphism φ : F 7→ E

is a K-homomorphism if φ is also a K-module homomorphism (that is, ∀k ∈ K,
φ(k) = k, or simply, φ|K , that is, φ when restricted to K, is the identity map of K.
In this case, we say that K is a fixed field of φ). When E = F , a map φ ∈ Aut(F )
is a K-automorphism of F if K is a fixed field of φ.

(3.2) The set of all K-automorphism of F forms a group under map composition.
This group will be denoted by AutK F , called the Galois group of F over K.

(3.3) Let K ≤ F be fields, f ∈ K[x], u ∈ F and φ ∈ AutK F . If f (u) = 0, then

f (φ(u)) = φ(f (u)) = φ(0) = 0, and so φ(u) is also a root of f (x) in F .

(3.4) Let K ≤ F be fields and F is algebraic over K. Two elements α, β ∈ F

are conjugate over K if irr(α, K) = irr(β, K).

(3.5) (Conjugate Isomorphism) Let α, β be algebraic elements of degree n over a

field F . The map
n−1 n−1
ci α i ) = ( ci β i ),
X X
φα,β (
i=0 i=0

defines an isomorphism of F (α) onto F (β) if and only if α abd β are conjugate over F .

Proof: Use the isomorphism and its inverse to show that irr(α, F )|irr(β, F ) and
irr(β, F )|irr(α, F ).
Conversely, one can show that φα,β is a bijective homomorphism.

(3.6) Let α be algebraic over F . Then

(i) Any isomorphism φ from F (α) onto subfield of F̄ maps α to β = φ(α) which
is conjugate to α over F .
(ii) For each conjugate β of α, there is only one isomorphism, namely φα,β , map-
ping F (α) onto F (β), mapping α to β, and fixing elements in F .

√
(3.7) Example of Computing Galois groups: Let K = Q and F = Q( 2).

9
 √ √ √
Let f (x) = irr( 2, Q). Then f (x) = x2 − 2. Note that both 2 and − 2 are roots
√
of f (x) and both are in F . By (3.3), any element φ ∈ AutK F must send 2 to
√ √
either 2 or to − 2, and by (3.5), the image of φ uniquely determines φ. Hence
|AutK F | = 2 and so AutK F ∼
= Z2 .

√ √
Similarly, when F = Q( 2, 3) and K = Q, then AutK F ∼
= Z2 × Z2 .

(3.8) Let K ≤ E ≤ F be fields, G = AutK F and H ≤ G. Each of the following

holds:
(3.8-1) H 0 = {v ∈ F : ∀φ ∈ H, φ(v) = v} is an field such that K ≤ H 0 ≤ F (called
the fixed field of H).
(3.8-2) E 0 = {φ ∈ AutK F : ∀u ∈ E, φ(u) = u} = AutE F ≤ AutK F .
(3.8-3) In particular, G0 = F , and F 0 = AutF F is the single element group

Proof: Verify the definitions.

(3.9) Let K ≤ F be fields and G = AutK F . If G0 = K, then F is a Galois ex-

tension (field) of K (we also say that F is Galois over K).

√
(3.10) Example: Q( 2) is Galois over Q.

4. The Fundamental Theorem

(4.1) Let F, K, L, M be fields such that K ≤ M ≤ F and K ≤ L ≤ F , G = AutK F ,

and H, J ≤ G be subgroups. Each holds:
(4.1-1) F 0 = 1G and K 0 = G.
(4.1-1)’ (1G )0 = F .
(4.1-2) L ≤ M =⇒ M 0 ≤ L0 .
(4.1-2)’ H ≤ J =⇒ J 0 ≤ H 0 .
(4.1-3) L ≤ L00 := (L0 )0 and H ≤ H 00 := (H 0 )0 .
(4.1-4) L0 = L000 and H 0 = H 000 .
Proof: Verify definitions. For (4.1-4), By (4.1-3), we have L ≤ L00 and so by (4.1-2)
(with M = L00 ), we have L000 ≤ L0 . Apply (4.1-3) with L0 replacing L to get L0 ≤ L000 .

10