JIA Jia

Notes on Algebra
December 16, 2018

Springer
Contents

1 Algebra-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Noetherian Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Artinian Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6 Tensor Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.7 Multi-linear Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Commutative Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 Exact Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Tensor Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Flatness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Limits, colimits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 Fields Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1

Algebra-1

1.1 Groups

Proposition 1.1. Let G be a finite group. Suppose that G has at most one subgroup of order d whenever d | |G|. Prove
that G is cyclic.

Proof. Denote n = |G|. If there is a subgroup H ≤ G with order d | n, all elements of G with order d must in H. Otherwise,
let g ∈ G \ H with o(g) = d, then hgi and H will be two distinct subgroups of G with order d, a contradiction.
P
So, G has at most ϕ(d) elements of order d for each d | n. By Euler’s identity d|n ϕ(d) = n, we know that G contains
exactly ϕ(d) elements of order d for each d | n. In particular, G has an element of order n. Thus G is cyclic. 

Proposition 1.2. Let P be a Sylow p-subgroup of a group G, and let N be a normal subgroup of G. Show that P ∩ N is
a Sylow p-subgroup of N , and that P N/N is a Sylow p-subgroup of G/N .

Proof. Suppose |G| = pa m with p - m.

By assumption |P | = pa . Since N C G, we have P ∩ N C P . Therefore |P ∩ N | = pr for some r ≥ 1.
The second isomorphism theorem implies that P N/N ∼
= P/(P ∩ N ). Thus

|N |/|P ∩ N | = |P N |/|P |.

It follows that [N : P ∩ N ] = [P N : P ] | [G : P ] = m. We deduce that r is the largest power of p in |N |. So P ∩ N is a

Sylow p-subgroup of N .
Meanwhile, |P N/N | = |P |/|P ∩ N | = ps for some s ≥ 1. Note that [G/N : P N/N ] = [G : P N ] | [G : P ] = m. Similarly,
we have P N/N is a Sylow p-subgroup of G/N . 

Proposition 1.3. If G is an Abelian p-group has a unique subgroup of order p, then G is cyclic.

Proof. Take g ∈ G with maximal order, say o(g) = pk .

Claim:hgi = G.
Otherwise, pick xhgi ∈ G/hgi with o(xhgi) = p. Then x ∈ G \ hgi and xp ∈ hgi. Hence, xp = g a for some a.
k k−1 k−1
But we know 1 = xp = (xp )p = (g a )p . So, o(g) = pk | apk−1 . Therefore, p | a.
Suppose that a = pb for some b. we have y = xg −b ∈ xhgi. Hence, y ∈
/ hgi and y p = (xg −b )p = xp g −bp = 1. Note that
k−1
hyi and hg p i are two distinct subgroups of G of order p, a contradiction. 
2 1 Algebra-1

Proposition 1.4. Let P be a Sylow p-subgroup of a finite group G, and let H be a subgroup of G containing NG (P ).
Prove that H is self-normalising, i.e. NG (H) = H. In particular, this shows that NG (P ) is self-normalising.

Proof. One direction is obvious, so we need to show that NG (H) ⊆ H.

Let g ∈ NG (H), then gP g −1 ≤ gHg −1 = H. So, gP g −1 is a Sylow p-subgroup of H. Since P is also a Sylow p-
subgroup of H, by Sylow’s 2nd Theorem there exists x ∈ H such that P = xgP g −1 x−1 . Therefore, xg ∈ NG (P ) ≤ H and
g ∈ x−1 H = H. 

Proposition 1.5. Let p be a prime number. Show that any non-Abelian group of order p3 has a centre of order p, and
that it has exactly p2 + p − 1 conjugacy classes.

Proof. Every p-group has a non-trivial center. If G/Z(G) is cyclic, then G is Abelian. So we have |Z(G)| = p.
Note that every element of Z(G) belongs to a conjugacy class consists of exactly itself. Let G act on itself by conjugation.
For every x ∈ G \ Z(G), |Stab(x)| is a power of p. But Stab(x) contains the center and x itself, so |Stab(x)| ≥ p + 1.
Since |Stab(y)| = p3 if and only if y ∈ Z(G), we have |Stab(x)| = p2 . It follosws from the Orbit-Stabilizer theorem that
|orbit of x| = p. Hence there are (p3 − p)/p = p2 − 1 conjugacy classes except Z(G). The total number is p2 + p − 1. 

Proposition 1.6. Let P be a p-subgroup of group G, Q a Sylow p-subgroup of G. Then P ∩ NG (Q) = P ∩ Q.

Proof. Note that P ∩ NG (Q) is a p-subgroup of NG (Q). However, Q is normal in NG (Q), so it is the only one Sylow
p-subgroup of NG (Q). It follows from Sylow’s 2nd Theorem that P ∩ NG (Q) ⊂ Q. Therefore P ∩ NG (Q) = P ∩ Q. 

Proposition 1.7. Prove that a group of order pqr, where p > q > r are primes with p, q 6≡ 1 (mod r), p 6≡ 1 (mod q), is
cyclic.

Proof. Claim: At least one of np and nq equals to 1.

Suppose np = 1, let P, Q, R be Sylow p, q, r-subgroups, respectively. Then P is normal in G. Note that P Q is a group
of order pq, therefore is cyclic. Similar for P R. We have CG (P ) ⊇ hP, Q, Ri, hence CG (P ) = G. In particular, we have
P C Z(G).
Now by 3rd Isomorphism Theorem, G/Z(G) ∼
= (G/P )/(Z(G)/P ). Since G/P has order qr, it is cyclic, so is the quotient
group (G/P )/(Z(G)/P ). It follows from that G/Z(G) is cyclic and G is Abelian.
Thus Q, R are both normal in G and G = P QR ∼
= Cpqr is cyclic. 

Proposition 1.8.
(a) A group of order 30 has a normal Sylow 5-subgroup.
(b) A group of order 60 is either simple or has a normal Sylow 5-subgroup.

Proof.

(a) Let |G| = 30. If n5 6= 1, then n5 = 6, and so NG (P5 ) = P5 . But then n3 = 1 and so P5 P3 is a normal subgroup of G, and
consequently P3 ⊆ NG (P5 ), a contradiction. (Or P5 is a normal Sylow 5-subgroup of P5 P3 and hence characteristic.
Now we will have P5 C G, a contradiction)
 1.2 Modules 3

(b) Let |G| = 60. If n5 6= 1, then n5 = 6 and |NG (P5 )| = 10. Let N be a non-trivial normal subgroup of G. If 5 | |N |, then
|N | ∈ {5, 10, 15, 20, 30}, and in all these cases N has a normal, and hence characteristic, Sylow 5-subgroup which is
then normal in G, contradicting n5 = 1. Thus 5 - |N |, and so 5 | |G/N |. Pulling back the Sylow 5-subgroup of G/N
gives a subgroup H of G of order 5|N |. If H = G, then |N | = 12 and N has a normal Sylow 2-subgroup or 3-subgroup,
which are charicteristic. Either |P2 | = 4 and P2 C G or |P3 | = 3 and P3 C G. Now either P5 C P5 P2 or P5 P3 . So either
P5 P2 ≤ NG (P5 ) or P5 P3 ≤ NG (P5 ), which implies that 20 | 10 or 15 | 10, absurd. So H  G and H has a normal
Sylow 5-subgroup P5 , so that NG (P5 ) ⊇ H. Since |NG (P5 )| = 10, this yields |N | = 2. But then G/N has a normal
Sylow 5-subgroup and hence exactly 4 elements of order 5, while G has 24 elements of order 5, a contradiction.

Proposition 1.9. If G is a finite group, then for any group element a, the elements in the conjugacy class of a are in
one-to-one correspondence with cosets of the centralizer CG (a).

Proof. This can be seen by observing that any two elements b and c belonging to the same coset (and hence, b = cz for
some z in the centralizer CG (a)) give rise to the same element when conjugating a : bab−1 = cza(cz)−1 = czaz −1 c−1 =
czz −1 ac−1 = cac−1 . The converse holds as well. 

1.2 Modules

Proposition 1.10. Let M be a free R-module with a basis B. If B is finite (as a set), then every basis of M is finite.

Proof. Let B and B 0 be bases of M , B = {b1 , b2 , · · · , bn }.

Since bi ∈ M = b0 ∈B 0 Rb0 , we know that for each i, there is a finite set Ci ⊂ B 0 such that bi ∈ c∈Ci Rc.
L P
Pn
However, M = i=1 Rbi ⊂ c∈Sn Ci Rc ⊂ b0 ∈B 0 Rb0 = M . So, they must be equal.
P P
i=1
Sn
If there is b0 ∈ B 0 \ i=1 Ci , then b0 = 1b0 = c∈Sn Ci rc c, contradicts the direct sum.
P

i=1

Proposition 1.11. Let M be a finitely generated R-module. Prove that if J(R)M = M , then M = 0.

Proof. Suppose M is not 0. Then M has a maximal submodule N as it is finitely generated. Notice that M/N is
simple and J(R) is the intersection of all the annihilators of simple R-modules, We have J(R)(M/N ) = 0. Therefore
J(R)M ⊂ N M , a contradiction. 

Proposition 1.12. Prove that R has IBN if and only if there does not exist invertible non-square matrix over R. (An
(n × m)-matrix M over R is invertible if and only if there exists an (m × n)-matrix N over R such that M N = In and
N M = Im .)

Proof. If there exists an (n × m)-matrix M and an (m × n)-matrix N such that M N = In and N M = Im , then we can
define the following two R-module homomorphisms:

φ : Rm → Rn , v t 7→ v t M t ;

ψ : Rn → Rm , ut 7→ ut N t .
4 1 Algebra-1

It is easy to see that φ and ψ are inverse to each other, so Rm ∼

= Rn as R-modules. So, if m 6= n, then R does not has
IBN.
ϕ
If R does not have IBN, then there exist two distinct integers m, n such that Rm ∼
= Rn as R-modules, then define M
be the matrix of ϕ under the canonical base, and N the matrix of ϕ−1 . Then M, N are invertible to each other.
Alternatively,given a free R-module B and assume that E = {b1 , · · · , bm } is a basis. Define F = M E, then F is also a
basis of B.(check that!) 

Proposition 1.13. Let M be an R-module. Show that the following statements are equivalent:
(a) M is Noetherian.
(b) Every submodule of M is finitely generated.
(c) Every non-empty collection of finitely generated submodules of M has a maximal element.

Proof. (a)⇒(b) If N is a non-finitely generated submodule of M , take n1 ∈ N , then Rn1  N . When find n1 , · · · , nk ∈
Pk Pk
N, i=1 Rni  N , so there exists nk+1 ∈ N \ i=1 Rni . So we constructed

k
X k+1
X
Rn1  Rn1 + Rn2  · · ·  Rni  Rni  · · ·
i=1 i=1

a strictly ascending chain of submodules of M , contradiction.

(b)⇒(c) Let Σ be a nonempty collection of finitely generated submodules and assume that Σ has no maximal elements.
Take Ni ∈ Σ, for the chain N1  · · ·  Nk , there exists Nk+1 ∈ Σ such that Nk+1 Ni . Then N1  · · · Nk  · · · is
S∞ S∞
a strictly ascending chain. Let N := i=1 Ni ≤ M , so N is finitely generated, say by n1 , · · · , nm . Each nj ∈ i=1 Ni , so
nj ∈ Nij for some ij .
S∞
Let k ≥ max{ij }, then nj ∈ Nk for all j. So we have Nk = i=1 Ni with every k ≥ max{ij }, contradicts Nk  Nk+1 
···.
(c)⇒(a) Suppose M is not Noetherian, then there exist Ni ≤ M such that N1  N2  N3  · · · . Take n1 ∈ N1 , n2 ∈
Pl
N2 \ N1 , · · · , nk ∈ Nk \ Nk−1 , define Nl0 = j=1 Rnj . Then {N10 , N20 , · · · } is a nonempty collection of finitely generated
submodules of M without any maximal elements. 

Proposition 1.14. Suppose that R is commutative, and let M be a Noetherian R-module. Show that R/Ann(M ) is a
Noetherian R-module.
Pk
Proof. M is Noetherian so it is finitely generated: M = i=1 Rmi . Since R is commutative, we have Ann(M ) =
Tk
i=1 Ann(mi ).

Now R/Ann(M ) is Noetherian as R/Ann(mi ) ∼

= Rmi ≤ M are Noetherian and by T06-Q6. 

1.3 Localization

Proposition 1.15. Let A be an integral domain with field of fractions K, and let Am denote the localization of A at a
maximal ideal m considered as a subring of K. Prove that

\
A= Am
m
 1.4 Noetherian Rings 5

where the intersection is taken over all maximal ideals m of A.

T
Proof. Since A is an integral domain, A ⊆ Am for all maximal ideal m. Hence A ⊆ m Am .
Conversely, pick z ∈ K(A) \ A and consider the ideal I := (A : z) = {x ∈ A : xz ∈ A}. Notice that 1 6∈ I, then I is
a proper ideal and hence contained in a maximal ideal m. Claim: z ∈
/ Am . Otherwise, there exists some s ∈
/ m such that
sz ∈ A. Then s ∈ I ⊆ m, absurd. 

1.4 Noetherian Rings

Proposition 1.16. Suppose A is a Noetherian ring, so is A[[x]].

Proof. If I is an ideal of A[[x]], we must show that I is finitely generated. We will inductively construct a sequence of
elements fi ∈ A[[x]] as follows.
Let fi have minimal degree among elements of I. Suppose that we have chosen f1 , · · · , fi , where fi has degree di and
leading coefficient ai . We then select fi+1 satisfying the following requirements:
(i) fi+1 ∈ I;
(ii) ai+1 6∈ {a1 , · · · , ai };
(iii) among all elements satisfying the first two conditions, f[ i + 1] has minimal degree.
The second condition forces the precedure to terminate in a finite number of steps. If stablization occers at step k, we
will show that I is generated by f1 , · · · , fk .
Let g = axd + · · · be an element of I of degree d and leading coefficient a. Then a ∈ (a1 , · · · , ak ).
Pk
Case 1: d ≥ dk . Since di ≤ di+1 for all i, we have d ≥ di for i = 1, · · · , k. Now a = i=1 ci0 ai with ci0 ∈ A. Define

k
X
g0 = ci0 xd−di fi ,
i=1

so that g0 has degree d and the leading coefficient a and consequently, g − g0 has degree greater that d. Having define
Pr Pr
g1 , · · · , gr ∈ (f1 , · · · , fk ) such that g − j=0 gj has degree greater that d + r, say g − j=0 gj = bxd+r+1 + · · · . Now
Pk
b ∈ (a1 , · · · , ak ) so b = i=1 ci,r+1 ai with ci,r+1 ∈ A. We define

k
X
gr+1 = ci,r+1 xd+r+1−di fi ,
i+1

Pr+1
so that g − j=1 gj has degree greater that d + r + 1. Thus

∞
X ∞ X
X k
g= gr = ci,r xd+r−di fi ,
r=0 r=0 i=1

and it follows upon reversing of the order of summation that g ∈ (f1 , · · · , fk ). (inner summation is finite, so for a given
integer l, there are only finitely many terms of the form exl )
Case 2: d < dk . As above, a ∈ (a1 , · · · , ak ), so there is a smallest n between 1 and k such that a ∈ (a1 , · · · , am ). It
Pm
follows that d ≥ dm . As in case 1, we have a = i=1 ci ai with ci ∈ A. Define
6 1 Algebra-1
m
X
h= ci xd−di fi ∈ (f1 , · · · , fk ) ⊆ I.
i=1

The leading coefficient of h is a, so the degree of g − h is greater that d. We replace g by g − h and repeat the procedure.
P
After at most dk − d iterations we produce an element g − hi in I of degree at least dk , with all hi ∈ (f1 , · · · , fk ). By
the analysis in case 1, g ∈ (f1 , · · · , fk ). 

1.5 Artinian Rings

Proposition 1.17. If R is Artinian, then J(R) is nilpotent.

Proof. Notice that

J ⊇ J2 ⊇ J3 ⊇ · · ·

Since R is Artinian, J n = J n+1 = · · · for some n ∈ Z≥1 .

If J n 6= {0}, define
X
K= Rx.
x∈R,
J n (Rx)={0}

If N ≤ R with J n N = {0}, then N ≤ K. So K is the maximal ideal killed by J n .

Since J n 6= {0}, K 6= R. Since R is Artinian, there exists a minimal ideal K 0 of R strictly containing K. Now K 0 /K
is simple, so J(K 0 /K) = {0}. From this we know JK 0 ⊂ K. Hence

J n K 0 = J n+1 K 0 = J n (JK 0 ) ⊂ J n K = {0}.

It follows that K 0 ⊂ K, a contradiction. 

Theorem 1.18 (Chinese Remainer Theorem). Let I1 , · · · , In be ideals of R such that Ii + Ij = R for any i 6= j. Then

n n
Ij ∼
\ Y
R/ = R/Ij .
j=1 j=1

Proof. Define
n
Y
φ : R −→ R/Ij ,
j=1
Tn
it is clear that ker φ = j=1 Ij .

Claim: for each i, there exist ai ∈ Ii and ci = b1 · · · bˆi · · · bn ∈

T
j6=i Ij where bj ∈ Ij such that ai + ci = 1R .
For any i 6= j there exist xj ∈ Ii and yj ∈ Ij such that xj +yj = 1R . So ci = y1 · · · yˆi · · · yn = (1−x1 ) · · · (1\
− xi ) · · · (1−
xn ) = 1 − ai for some ai ∈ Ii .
Now φ(ci ) = (c1 + I1 , · · · , ci + Ii , · · · , cn + In ) = (0 + I1 , · · · , 1 − ai + Ii , · · · , 0 + In ) = (0 + I1 , · · · , 1 + Ii , · · · , 0 + In ).
So φ is surjective. 

Proposition 1.19. Let I be a non-finitely generated ideal of A such that any ideal J of A with I ( J is finitely generated.
Show that I is prime.
 1.7 Multi-linear Algebra 7

Proof. I 6= A since A is finitely generated. Let x, y ∈ A with xy ∈ I but x 6∈ I. Since I + (x) ) I we know I + (x) is
finitely generated, say by {ai + bi x}ni=1 with ai ∈ I and bi ∈ A.
Take c ∈ I ( I + (x), then
n
X
c= di (ai + bi x).
i=1
Pn Pn
So we have ( i=1 di bi ) x = c − i=1 di ai ∈ I. Since c is arbitrary, we have

n
X
I= Aai + (I : x) x.
i=1

From this we deduce that (I : x) is not finitely generated. However I ⊆ (I : x). So I = (I : x) and y ∈ (I : x) = I. 

1.6 Tensor Product

Remark 1.20. HomS (R NS , PS ) has a natural right R-module structure.

Remark 1.21. Let A be a commutative ring with 1 and R, S be A-algebras. Let M be a left R-module, N be a left S-
module. EndA (M ) = {φ ∈ EndZ (M ) : φ(am) = aφ(m) for every a ∈ A, m ∈ M } is a unital subring of EndZ (M ). The
A-module structure of EndA (M ) is compatible with its ring structure, so it is an A-algebra. Then M is a left R-module
is equivalent to there exists a unital A-algebra homomorphism ΦR : R → EndA (M ), N is a left S-module is equivalent to
there exists a unital A-algebra homomorphism ΦS : S → EndA (N ).

ΦR ⊗A ΦS : R ⊗ S EndA (M ) ⊗ EndA (N )

EndA (M ⊗A N )

1.7 Multi-linear Algebra

Remark 1.22 (Lemma 10.23). H = {1Sn , (ij)} is a subgroup of Sn , say Γ is a set of left coset representation of H in Sn .
Then
a a [a
Sn = gH = g g(ij).
g∈Γ g∈Γ g∈Γ

If mi = mj , then
8 1 Algebra-1

m1 ∧ · · · ∧ mn =  · (m1 ⊗ · · · ⊗ mn )
X
=( sgn(σ)σ) · (m1 ⊗ · · · ⊗ mn )
σ∈Sn
   
X X
=  sgn(g)g  +  sgn(g(ij))g(ij) · (m1 ⊗ · · · ⊗ mn )
g∈Γ g∈Γ
   
X X
= sgn(g)g  · (m1 ⊗ · · · ⊗ mn ) +  sgn(g(ij))g(ij) · (m1 ⊗ · · · ⊗ mn )
g∈Γ g∈Γ
   
X X
= sgn(g)g  · (m1 ⊗ · · · ⊗ mn ) +  (−1) sgn(g)g  · (m1 ⊗ · · · ⊗ mn )
g∈Γ g∈Γ

= 0.

Remark 1.23 (Lemma 10.23).

mσ−1 (1) ∧ · · · ∧ mσ−1 (n) =  · (mσ−1 (1) ⊗ · · · ⊗ mσ−1 (n) )

=  · (σ(m1 ⊗ · · · ⊗ mn ))

= ( · σ) · (m1 ⊗ · · · ⊗ mn )

= (sgn(σ)) · (m1 ⊗ · · · ⊗ mn )

= sgn(σ)( · (m1 ⊗ · · · ⊗ mn ))

= sgn(σ)(m1 ∧ · · · ∧ mn ).

Remark 1.24 (Prop 11.10).

!  

λσ σ(a1), y = (a1)
X X X
λσ σ, y  7−→ ( λσ σ)ay + IS (M );
σ 16=σ σ

(a1,
X X
λσ σy) 7−→ a( λσ σ)y + IS (M ).
σ σ

Remark 1.25. Prove or disprove that if 2 = 1A + 1A is a unit in A,

(M )n ∼
^
= ASn ⊗ASn T n (M ).

Proposition 1.26. Let A be a commutative ring with 1. Let M be a free A-module with basis {mj | j ∈ J}. Let n ∈ Z+ .
Recall that IS (M ) denotes the two-sided ideal of the tensor algebra T (M ) of M generated by {m⊗m0 −m0 ⊗m | m, m0 ∈
M }, and that
ntimes
z }| {
n
T (M ) = M ⊗ · · · ⊗ M

denotes the n-th tensor power of M .

(a) Prove that IS (M ) ∩ T n (M ) is free as an A-module.
 1.7 Multi-linear Algebra 9

(b) Hence, or otherwise, show that the n-th symmetric power of M , denoted S n (M ), is free as an A-module, and write
down a basis for S n (M )

Proof. For j = {j1 , · · · , jn } ∈ J n , let mj = mj1 ⊗ · · · ⊗ mjn . Let C ⊆ J n consists of {j1 , · · · , jn } with j1 ≤ j2 ≤ · · · ≤ jn .
Then we claim: IS (M ) ∩ T n (M ) is free with basis

{mj − σ · mj | j ∈ C, σ ∈ Sn and σ · j 6= j}

Since σ · mj with j ∈ C, σ ∈ Sn and σ · j 6= j are linearly independent in T n (M ), the asserted basis is also linearly
independent. They are in IS (M ) by Lemma 11.9. 
2

Commutative Algebra

Proposition 2.1 (Fields and algebraic algebras). Let A be an algebra over a field K.

1. If A is an integral domain and algebraic over K, then A is a field.

2. If A is a field and is contained in an affine K-domain, then A is algebraic.

Proof.

1. We need to show that every a ∈ A \ {0} is invertible in A. For this, it suffies to show that K[a] is a field. We may
therefore assume that A = K[a]. With x an indeterminate, let I ⊂ K[x] be the kernel of the map K[x] → A, f 7→ f (a).
Then A ∼
= K[x]/I. Since A is an integral domain, I is a prime ideal, and since a is algebraic over K, I is nonzero.
Since K[x] is a principal ideal domain, it follows that I = (f ) with f ∈ K[x] irreducible, so I is a maximal ideal. It
follows that A ∼
= K[x]/I is a field.
2. By way of contradiction, assume that A has an element a1 that is not algebraic. By hypothesis, A is contained in an
affine K-domain B = K[a1 , · · · , an ] (we may include a1 in the set of generators). We can reorder a2 , · · · , an in such a
way that {a1 , · · · , ar } forms a maximal K-algebraically independent subset of {a1 , · · · , an }. Then the field of fractions
Quot(B) of B is a finite field extension of the subfield L := K(a1 , · · · , ar ). For b ∈ Quot(B), multiplication by b gives
an L-linear endomorphism of Quot(B). Choosing an L-basis of Quot(B), we obtain a map φ : Quot(B) → Lm×m
assigning to each b ∈ Quot(B) the representation matrix of this endomorphism. Let g ∈ K[a1 , · · · , ar ] \ 0 be a common
denominator of all the matrix entries of all φ(ai ), i = 1, · · · , n. So φ(ai ) ∈ K[a1 , · · · , ar , g −1 ]m×m for all i. Since φ
preserves addition and multiplication, we obtain

φ(B) ⊆ K[a1 , · · · , ar , g −1 ]m×m .

K[a1 , · · · , ar ] is isomorphic to a polynomial ring and therefore factorial . Take a factorization of g, and let p1 , · · · , pk
be those irreducible factors of g that happen to lie in K[a1 ]. Let p ∈ K[a1 ] be an arbitrary irreducible element. Then
p−1 ∈ A ⊆ B since K[a1 ] ⊂ A and A is a field. Applying φ to p−1 yields a diagonal matrix with all entries equal
to p−1 , so there exists a nonnegative integer s and an f ∈ K[a1 , · · · , ar ] with p−1 = g −s · f , so g s = p · f . By the
irreducibility of p, it follows that p is a K-multiple of one of the pi . Since this holds for all irreducible elements
Qk
p ∈ K[a1 ], every element from K[a1 ] \ K is divisible by at least one of the pi . But none of the pi divides i=1 pi + 1.
This is a contradiction, so all elements of A are algebraic.


12 2 Commutative Algebra

2.1 Exact Sequences

Proposition 2.2. Let R be a ring.

1. Let 0 → L → Rn → M → 0 be an exact sequence. Then M is finitely generated if and only if L is finitely generated.
α β
2. Let 0 → L −
→M −
→ N → 0 be a short exact sequence.
a) If L is finitely generated and M is finitely presented then N is finitely presented.
b) If M is finitely generated and N is finitely presented, then L is finitely generated.
c) If L and N are finitely presented, then M is finitely presented.

Proof.

1. Assume M is finitely presented; say RI → Rm → M → 0 is a finite presentation. Let L0 be the image of Rl . Then
L0 ⊕ Rn ' L ⊕ Rm by Schanuel’s Lemma. Hence L is a quotient of Rl ⊕ Rn . Thus L is finitely generated.
Conversely, assume L is generated by l elements. They yield a surjection Rl  L. It yields a sequence Rl → Rn →
M → 0. The latter is, plainly, exact. Thus M is finitely presented.
2. a) Let µ : Rm → M any surjection. Set ν := βµ, set K := Kerν, and set λ := µ|K . Then the following diagram is
commutative:
ν
0 K Rm N 0
λ µ 1N
α β
0 L M N 0
The Snake Lemma yields an isomorphism ker λ → ker µ. But ker µ is finitely generated by 1. So ker λ is finitely
generated. Also, the Snake Lemma implies Coker λ = 0 as Coker µ = 0; so 0 → ker λ → K → L → 0 is exact.
Hence K is finitely generated. Thus N is finitely presented by 1.
b) Say M is generated by m elements. They yield a surjection µ : Rm  M . As in (a), µ induces the following
commutative diagram, with λ surjective:
0 K Rm ν N 0
λ 1N
α β
0 L M N 0
By 1, K is finitely generated. Thus L is too, as λ is surjective.
c) Let λ : Rl → L and ν : Rn  N any surjections. Define γ : Rl → M by γ := αλ. Note Rn is projective, and define
δ : Rn → M by lifting ν along β. Define µ : Rl ⊕ Rn → M by µ := γ + δ. Then the following diagram is, plainly,
commutative, where ι := ιRl and π := πRn :
ι π
0 Rl Rl ⊕ Rn Rn 0
λ µ ν
α β
0 L M N 0
Since λ and ν are surjective, the Snake Lemma yields an exact sequence

0 → ker λ → ker µ → ker ν → 0,

and implies Coker µ = 0. Also, ker λ and ker ν are finitely generated by 1. So ker µ is finitely generated. Thus M
is finitely presented by 1.


 2.2 Tensor Products 13

2.2 Tensor Products

Proposition 2.3. Let R be a ring, R0 an R-algebra, M, N two R0 -modules. Show there is a canonical R-linear map
τ : M ⊗R N → M ⊗R 0 N .
Let K ⊂ M ⊗R N denote the R-submodule generated by all the differences (x0 , m) ⊗ n − m ⊗ (x0 , n) for x0 ∈ R0 and
m ∈ M and n ∈ N . Show K is equal to ker(τ ), and τ is surjective.

Proof. The canonical map β 0 : M × N → M ⊗R0 N is R0 -bilinear, so R-bilinear. Hence, it factors: β 0 = τ β where
β : M × N → M ⊗R N is the canonical map and τ is the desired map.
Set Q := (M ⊗R N )/K. Then τ factors through a map τ 0 : Q → M ⊗R0 N since each generator (x0 , m) ⊗ n − m ⊗ (x0 , n)
of K maps to 0 in M ⊗R0 N .
There is an R0 -structure on M ⊗R N with y 0 (m ⊗ n) = m ⊗ (y 0 , n), and so, another one with y 0 (m ⊗ n) = (y 0 , m) ⊗ n.
Clearly, K is a submodule for each structure, so Q is too. But on Q the two structures coincide. Further, the canonical map
M × N → Q is R0 -bilinear. Hence the latter factors through M ⊗R0 N , furnishing an inverse to τ 0 . So τ 0 : Q → M ⊗R0 N .
Hence ker(τ ) is equal to K, and τ is surjective. 

Remark 2.4. If a R-linear map δ : M ⊗R N → P satisfies that δ((x0 m) ⊗ n) = δ(m ⊗ (x0 n)), then δ induced a R0 -linear
map δ 0 : M ⊗R0 N → P .

Property 2.5. 1. Let R be a rings, M and N modules.

a) Then the switch map (m, n) → (n, m) induces an isomorphism

M ⊗R N = N ⊗R M.(commutative law)

b) Then multiplication of R on M induces an isomorphism

R ⊗R M = M.(unitary law)

2. Let R and R0 be rings, M an R-module, P an R0 -module, N an (R, R0 )-bimodule. Then there are two canonical
(R,R0 )-isomorphisms:

M ⊗R (N ⊗R0 P ) = (M ⊗R N ) ⊗R0 P, (associative law)

HomR0 (M ⊗R N, P ) = HomR (M, HomR0 (N, P ).(adjoint associativity)

3. Let R be a ring, and R0 an algebra. First, let M be an R-module, and P an R0 -module. Then there are two canonical
R0 -isomorphisms:

(M ⊗R R0 ) ⊗R0 P = M ⊗R P, (cancellation law)

HomR0 (M ⊗R R0 , P ) = HomR (M, P ).(left adjoint)

Instead, let M be an R0 -module, and P an R-module. Then there is a canonical R0 -isomorphism:

14 2 Commutative Algebra

HomR (M, P ) = HomR0 (M, HomR (R0 , P )).(right adjoint)

In other words, • ⊗R R0 is the left adjoint of restriction of scalars from R0 to R, and HomR (R0 , •) is its right adjoint.

Remark 2.6. Let R be a ring, R0 a R-algebra. Restriction of scalars from R0 to R can be realize as the functor P 7→R
HomR0 (R0 , P ) for any R0 -module P . Here, we view HomR0 (R0 , P ) as a R-module.

2.3 Flatness

Remark 2.7. 1. A functor F is exact if it preserves exact sequences;

2. A functor F is faithful if the preimage of 0 map is 0;
3. In R-module category, an exact functor F is faithful if the preimage of 0 module is 0; if the image of every simple
R-module is non-zero; if the preimage of exact sequence is exact.

2.4 Limits, colimits

Proposition 2.8. Assume that (πi : L → Bi )i is a limit cone then two morphisms f, g : A → L are equal if and only if
for each i we have πi ◦ f = πi ◦ g.

Proof.


2.5 Localization

Proposition 2.9. Let A be a ring, S ⊂ T ⊂ U multiplicative subsets.

1 Show there is a unique A-algebra map ϕST : S −1 A → T −1 A and ϕTU ϕST = ϕSU .
S
2 Let Λ be a set, Sλ ⊂ S a multiplicative subset for all λ ∈ Λ. Assume Sλ = S. Assume given λ, µ ∈ Λ, there is ν such
that Sλ , Sµ ⊂ Sν . Order Λ by inclusion: λ ≤ µ if Sλ ⊂ Sµ . Using (1), show lim Sλ−1 A = S −1 A.
−→

Proof.

1 Note ϕT (S) ⊂ ϕT (T ) ⊂ (T −1 A)× . So the universal property yields a unique A-algebra map ϕST : S −1 A → T −1 A. By
uniqueness, ϕTU ϕST = ϕSU . Thus (1) holds.
2 Notice Λ is directed. Given λ ≤ µ, set αµλ := ϕSSλµ . Then ανµ αµλ = ανλ if λ ≤ µ ≤ ν. Thus lim Sλ−1 A exists as a filtered
−→
direct limit of A-algebras.
Given λ, set βλ := ϕSSλ . Then βµ αµλ = βλ . So the βλ induce an A-algebra map β : lim Sλ−1 A → S −1 A with βλ = βαλ
−→
where αλ is the insertion of Sλ−1 A.
 2.6 Fields Theory 15
αλ

αλ αµ
Sλ−1 A lim Sλ−1 A
µ
Sµ−1 A
−→
βλ
βµ
β

S −1 A

Take x ∈ ker β. There are λ and xλ /sλ ∈ Sλ−1 A such that αλ (xλ /sλ ) = x. Then βλ (xλ /sλ ) = 0. So there is s ∈ S with
sxλ = 0. But s ∈ Sµ for some µ ≥ λ. Hence αµλ (xλ /sλ ) = 0. So x = αµ αµλ (xλ /sλ ) = 0. Thus β is injective.
As to surjectivity, take x/s ∈ S −1 A. Then s ∈ Sλ for some λ, so x/s ∈ Sλ A. Hence βλ (x/s) = x/s. Thus β is surjective,
so an isomorphism. Thus (2) holds.

2.6 Fields Theory

Let L be an algebraic extension of a field K. An element α of L is a primitive element of L/K if L = K(α).

Theorem 2.10 (The theorem of the primitive element). Every finite separable extension L of K has a primitive
element (and hence every such extension L is a simple extension).

Proof. We shall give the proof only in the case in which K has infinitely many elements. If K is a finite field, then also L
is a finite field, and in that case we know every nonzero element of L is the power of a single element θ. This element θ is
then a primitive element.
Let L = K(β1 , · · · , βn ). We adjoin to L n + 1 “indeterminants” X, X1 , · · · , Xn , that is, we consider the polynomial
ring L[X, X1 , · · · , Xn ] and its quotient field L(X, X1 , · · · , Xn ). We set K ∗ = K(X1 , · · · , Xn ), L∗ = L(X1 , · · · , Xn ). We
have then L∗ = K ∗ (β1 , · · · , βn ), and L∗ is a finite algebraic separable extension of K ∗ since the βi , being separable over
K, are also separable over K ∗ . We consider in L∗ the element

α∗ = X1 β1 + · · · + Xn βn . (2.1)

Let F (X) be the minimal polynomial of α∗ in K ∗ [X]. The coefficients of F (X) are rational functions of X1 , · · · , Xn
with coefficients in K. Let g(X1 , · · · , Xn ) be a common denominator of these rational functions, where g(X1 , · · · , Xn ) is
then an element in K[X1 , · · · , Xn ]. Then g(X1 , · · · , Xn )F (X) = f (X, X1 , · · · , Xn ) ∈ K[X, X1 , · · · , Xn ], and we have

f (α∗ , X1 , · · · , Xn ) = 0. (2.2)

Let
G(X1 , · · · , Xn ) = f (X1 β1 + · · · + Xn βn , X1 , · · · , Xn ). (2.3)

Then G(X1 , · · · , Xn ) is a polynomial in X1 , · · · , Xn , with coefficients in L, and we have, by (2.2): G(X1 , · · · , Xn ) ≡ 0.

Therefore also the partial derivatives ∂G/∂Xi , i = 1, · · · , n, are all zero. By (2.3), we have then:
16 2 Commutative Algebra

βi f 0 (α∗ , X1 , · · · , Xn ) + fi (α∗ , X1 , · · · , Xn ) = 0, (2.4)

where

∂f (X, X1 , · · · , Xn )
f 0 (X, X1 , · · · , Xn ) = ,
∂X
∂f (X, X1 , · · · , Xn )
fi (X, X1 , · · · , Xn ) = ·
∂Xi

The left-hand side in each of the equations (2.4) is, by (2.1), a polynomial in L[X1 , · · · , Xn ], and hence is the zero
polynomial. Consequently, the equations (2.4) remain valid if we substitude for X1 , · · · , Xn , any elements of K. On the
other hand, we have f 0 (α∗ , X1 , · · · , Xn ) = g(X1 , · · · , Xn )F 0 (X), and hence f 0 (α∗ , X1 , · · · , Xn ) 6= 0, since α∗ is separable
over K ∗ and therefore F 0 (α∗ ) 6= 0. Hence f 0 (α∗ , X1 , · · · , Xn ) is a nonzero polynomial in L[X1 , · · · , Xn ]. Since K ⊆ L and
K is an infinite field, we can find elements c1 , · · · , cn in K such that (c1 , · · · , cn ) is not a zero of that polynomial. We have
then, setting
α = c1 β1 + · · · + cn βn ,

that
f 0 (α, c1 , · · · , cn ) 6= 0 (2.5)

and
βi f 0 (α, c1 , · · · , cn ) + fi (α, c1 , · · · , cn ) = 0, i = 1, · · · , n. (2.6)

Equation (2.6) and inequality (2.5) imply that αi ∈ K(α), and since α ∈ L, it follows that L = K(α). This completes the
proof of the theorem.