MTH 302: Modules

Semester 2, 2013-2014

Dr. Prahlad Vaidyanathan

Contents
Finite Abelian Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
I. Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1. Definition and Examples . . . . . . . . . . . . . . . . . . . . . . . 3
2. Ideals and Quotient Rings . . . . . . . . . . . . . . . . . . . . . . 5
3. Prime and Maximal Ideals . . . . . . . . . . . . . . . . . . . . . . 6
4. The Chinese Remainder Theorem . . . . . . . . . . . . . . . . . . 7
5. Unique Factorization Domains . . . . . . . . . . . . . . . . . . . . 8
6. Principal Ideal Domains and Euclidean Domains . . . . . . . . . . 8
II. Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1. Definition and Examples . . . . . . . . . . . . . . . . . . . . . . . 9
2. Homomorphisms and Quotient Modules . . . . . . . . . . . . . . 10
3. Direct Sums of Modules . . . . . . . . . . . . . . . . . . . . . . . 11
4. Finitely Generated Modules . . . . . . . . . . . . . . . . . . . . . 11
III. Finitely Generated Modules over a PID . . . . . . . . . . . . . . . . . . . 14
1. Free Modules over a PID . . . . . . . . . . . . . . . . . . . . . . . 14
2. Torsion Modules over a PID - I . . . . . . . . . . . . . . . . . . . 15
3. Torsion Modules over a PID - II . . . . . . . . . . . . . . . . . . . 17
4. Finite Abelian Groups . . . . . . . . . . . . . . . . . . . . . . . . 18
5. Rational Canonical Form . . . . . . . . . . . . . . . . . . . . . . . 18
6. Cayley-Hamilton Theorem . . . . . . . . . . . . . . . . . . . . . . 21
7. Jordan Canonical Form . . . . . . . . . . . . . . . . . . . . . . . . 22
IV. Introduction to Commutative Algebra . . . . . . . . . . . . . . . . . . . . 24
1. Hom and Direct Sums . . . . . . . . . . . . . . . . . . . . . . . . 24
2. Exact Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3. Projective Modules . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4. Noetherian Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5. Noetherian Modules . . . . . . . . . . . . . . . . . . . . . . . . . 29
6. Artinian Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7. Length of a module . . . . . . . . . . . . . . . . . . . . . . . . . . 30
V. Tensor Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1. Finite Dimensional Vector Spaces . . . . . . . . . . . . . . . . . . 32
2. Tensor Product of Modules . . . . . . . . . . . . . . . . . . . . . 33
VI. Instructor Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2
Finite Abelian Groups
(See [Norman, §1])
0.1. Fundamental Theorem of finite abelian groups
0.2. Examples :
(i) If G = hgi is cyclic, then G ∼ = Z/mZ for some m ∈ Z
(ii) If G = {g1 , g2 , . . . , gn } finite abelian, then G ∼
= Zn /K for some subgroup
K < Zn
0.3. Definition :
(i) Z-basis for Zn .
(ii) Notation : K = hv1 , v2 , . . . , vt i
(iii) An invertible matrix P ∈ Mn (Z)
0.4. Examples :
(i) If K = h(2, 0), (0, 4i, then G = Z2 /K ∼= Z2 × Z4
 
4 6
(ii) If K 0 = h(4, 6), (8, 10)i, then write A = , then A is similar to D =
  8 10
2 0
, and so Z2 /K 0 ∼ = Z2 × Z4 as well.
0 4
0.5. Summary :
(i) Want to analyse a finite abelian group G, then write G ∼
= Zn /K for some
subgroup K < Zn
(ii) Find a Z-basis {v1 , v2 , . . . , vt } for K (Fact: Such a basis always exists with
t ≤ n)
 
v1
 v2  −1
(iii) Write A =  . . ., then find invertible matrices P, T such that P AT
 is
vt
diagonal (Fact: Such matrices P and T always exist)
(iv) Use the diagonal matrix to express G as a product of cyclic groups. The
diagonal matrix is called the Smith Normal form of A.

(End of Day 1)

I. Rings
1. Definition and Examples
1.1. Definition of a ring
1.2. Examples :

3
 (i) Z, Q, R, C. N is not a ring.
(ii) 2Z
(iii) Zn = Z/nZ
(iv) If R is a ring, then Mn (R) is a ring
(v) Z[i]
(vi) C[0, 1]
(vii) If R is a ring, then R[x] is a ring.
1.3. Definition :
(i) Multiplicative identity 1 = 1R . Note: If R has a multiplicative identity, then
it is unique.
(ii) Commutative ring
(iii) Division ring
(iv) Field
(v) Zero divisor
(vi) Integral domain
1.4. Examples :
(i) Q, R, C are fields. Z is not.
(ii) Zn is a field iff n is prime (See MTH 301.V.1.4)
(iii) M2 (R) has zero divisors.
(iv) Any finite integral domain is a field.
(v) If R is an integral domain, then R[x] is an integral domain.
1.5. Theorem: Let R be a ring, a, b ∈ R, then
(i) 0 · a = a · 0 = 0
(ii) (−a)b = a(−b) = −(ab)
(iii) (−a)(−b) = ab
(iv) (na)b = a(nb) = n(ab) ∀n ∈ Z
1.6. Definition of subring.
1.7. Definition of ring homomomorphism, isomorphism.
1.8. Examples :
(i) The quotient map Z → Zn
(ii) z 7→ z from C to C
(iii) f 7→ f (0) from C[0, 1] to C
(iv) x 7→ 2x from Z to Z is not a homomorphism.
1.9. Lemma: Let ϕ : R → R0 be a ring homomorphism, then

4
 (i) ϕ(0R ) = OR0
(ii) ϕ(−a) = −ϕ(a)
1.10. Definition :
(i) Kernel of a homomorphism
(ii) Image of a homomorphism
1.11. Proposition : ϕ is injective iff ker(ϕ) = {0}

2. Ideals and Quotient Rings

2.1. Definition of ideal
2.2. Examples :
(i) nZ C Z
(ii) {f ∈ C[0, 1] : f (0) = 0} C C[0, 1]
(iii) If I C R, then Mn (I) C Mn (R). Proof that the converse is true in a commu-
tative ring with 1.
(iv) If ϕ : R → R0 is a homomorphism, then ker(ϕ) C R
(End of Day 2)
2.3. Theorem: Let R be a ring, and I C R, then

R/I := {a + I : a ∈ R}

is a ring under the operations

(a + I) + (b + I) := (a + b) + I and (a + I)(b + I) := (ab) + I

Furthermore, the function π : R → R/I given by

π(a) := a + I

is a surjective ring homomorphism and ker(π) = I

2.4. (First Isomorphism Theorem): Let ϕ : R → R0 be a ring homomorphism, then
(i) ker(ϕ) C R
(ii) R/ ker(ϕ) ∼
= Im(ϕ)
In particular, if ϕ is surjective, then R/ ker(ϕ) ∼
= R0
2.5. Examples:
(i) If R = C[0, 1] and I = {f ∈ R : f (0) = 0}, then R/I ∼
=C
(ii) If I C R, then Mn (R)/Mn (I) ∼
= Mn (R/I)
(iii) If R is any ring, and I = (x) := {xf (x) : f (x) ∈ R[x]} C R[x], then R[x]/I ∼
=
R.

5
 2.6. (Correspondence theorem): If R is a ring, and I C R, then there is a 1-1 correspon-
dence between ideals in R/I and ideals in R containing I
2.7. Example : Ideals in Zn ↔ { divisors of n}
2.8. Lemma : Let R be a ring. Let {Sλ : λ ∈ Λ} be a (possibly uncountable) collection
of ideals in R. Then \
I := Sλ
λ∈Λ

is an ideal in R
2.9. Definition : Let R be a ring, and X ⊂ R, then
\
(X) := {I C R : X ⊂ I}

is called the ideal generated by X. If X = {a}, we write (a) = ({a}), and we call
(a) the principal ideal generated by a.
2.10. Lemma: Let R be a commutative ring with 1 ∈ R, and let X ⊂ R, then
∞ X
[ k
(X) = { ri xi : ri ∈ R, xi ∈ X}
k=1 i=1

ie. (X) consists of all finite sums of elements of the form rx where r ∈ R, x ∈ X.
In particular,
(a) = {ra : r ∈ R}

2.11. Example : If m, n ∈ Z, then ({m, n}) = (gcd(m, n)) =: (m, n)

(End of Day 3)

3. Prime and Maximal Ideals

Let R be a commutative ring with 1 ∈ R

3.1. Definition of Maximal ideal

3.2. Theorem: R is a field iff R has no non-trivial ideals
3.3. Theorem: I C R is a maximal ideal iff R/I is a field
3.4. Examples:
(i) If n ∈ Z is prime iff nZ C Z is a maximal ideal.
(ii) I := {f ∈ C[0, 1] : f (0) = 0} C C[0, 1] is a maximal ideal.
(iii) (x) C F [x] is a maximal ideal if F is a field.
(iv) (x) C Z[x] is not a maximal ideal.
3.5. Definition of prime ideal
3.6. Theorem: I C R is a prime ideal iff R/I is an integral domain.

6
3.7. Example : nZ C Z is prime iff n is a prime number.
3.8. Corollaries/Examples :
(i) (x) C Z[x] is a prime ideal.
(ii) If I C R is a maximal ideal, then I is a prime ideal (3.3+3.7)
(iii) Let ϕ : R → R0 be a surjective homomorphism, then ker(ϕ) C R is prime
(maximal) if R0 is an integral domain (field)

4. The Chinese Remainder Theorem

Let R be a commutative ring with 1 ∈ R
4.1. Definition :
(i) Sum of two ideals
(ii) Product of two ideals
(iii) Comaximal ideals
4.2. Lemma: Let I, J C R be comaximal ideals, then IJ = I ∩ J
4.3. (Chinese Remainder Theorem): Let I, J C R and define
ϕ : R → R/I × R/J, given by a 7→ (a + I, a + J)
(i) ϕ is a homomorphism and ker(ϕ) = I ∩ J
(ii) If I + J = R, then ϕ is surjective, and ker(ϕ) = IJ. In that case,
R/IJ ∼ = R/I × R/J
(End of Day 4)
4.4. Corollary: If I1 , I2 , . . . , In C R are such that for all i 6= j, one has Ii + Ij = R, then
R/(I1 I2 . . . In ) ∼
= R/I1 × R/I2 × . . . × R/In
4.5. Example :
(i) mZ, nZ C Z are comaximal iff (m, n) = 1. In that case, Znm ∼
= Zm × Zn
(ii) In particular, if n = pk11 pk22 . . . pkt t , then
Zn ∼ = Z k1 × Z p1
k
p2 2
× . . . × Zpkt
t

4.6. Definition :
(i) Unit of a ring R
(ii) The group of units R∗
4.7. Lemma: (R1 × R2 )∗ ∼= R1∗ × R2∗
4.8. Corollary: If n = pk11 pk22 . . . pkt t , then
Z∗n ∼ = Z∗pk1 × Z∗pk2 × . . . × Z∗pkt
1 2 t

Hence
ϕ(n) = (pk11 − pk11 −1 )(pk22 − pk22 −1 ) . . . (pkt t − pkt t −1 )

7
5. Unique Factorization Domains
Let R be an integral domain with 1 ∈ R

5.1. Definition : Let a, b ∈ R

(i) a | b (a divides b)
(ii) a ∼ b (a and b are associates)
5.2. Lemma :
(i) a | b iff (b) ⊂ (a)
(ii) a ∼ b iff (b) = (a). In particular, if u is a unit, then (u) = R
(iii) The relation ∼ is an equivalence relation.
5.3. Definition : Irreducible element
5.4. Remark :
(i) If r, s ∈ R such that r is irreducible and r ∼ s, then s is irreducible in R
(ii) Irreducibility depends on the ambient ring. 2 ∈ Z is irreducible, but 2 ∈ Q is
a unit.
5.5. Definition : R is a UFD iff R satisfies
(UF1) Every element can be written as a product of irreducibles and units.
(UF2) The above decomposition is unique upto a change of order.
5.6. Examples :
(i) Every field is a UFD
(ii) Z is a UFD. In fact, every Euclidean domain is a UFD
(iii) Z[x] is a UFD (MTH 301: §V I.4)
√ √ √
(iv) Z[ −5] is not a UFD because 6 = 2 · 3 = (1 + −5)(1 − −5) (with proof)
5.7. Definition : Prime element p ∈ R.
√
5.8. Lemma: Every prime is irreducible. Note: 2 ∈ Z[ −5] is irreducible, but not
prime.
(End of Day 5)
5.9. Theorem: R is a UFD iff R satisfies (UF1) and
(UF2’) Every irreducible is prime.

6. Principal Ideal Domains and Euclidean Domains

6.1. Definition
(i) Euclidean Domain
(ii) PID
6.2. Lemma: Every Euclidean domain is a PID

8
6.3. Examples :
(i) Every field is a Euclidean domain with d ≡ 0
(ii) Z is a Euclidean domain with d(n) = |n|
Z[i] is Euclidean domain with d(a + bi) = a2 + b2
(iii) If F is a field, F [x] is a Euclidean domain with d(f (x)) = deg(f (x))
(iv) Z[x] is not a PID (See MTH 301, Example § VI.4.3)
√
(v) If ω = (1 + −19)/2, then Z[w] is a PID, but not a Euclidean domain (proof
omitted)
6.4. Lemma: Every PID satisfies (UF1)
(End of Day 6)
6.5. Definition: Greatest Common Divisor (GCD)
6.6. Lemma: Let R be a PID, and a, b ∈ R. Then
(i) (a, b) exists in R
(ii) ∃x, y ∈ R such that (a, b) = ax + by
6.7. Theorem: Every PID is a UFD.

II. Modules
1. Definition and Examples
Let R be a commutative ring with 1 ∈ R
1.1. Definition of module M over R
1.2. Examples :
(i) R as a module over itself
(ii) A vector space over a field k is a k-module.
(iii) An abelian group is a Z-module. In fact, every Z-module is nothing but an
abelian group.
(iv) A linear transformation T : V → V gives V the structure of a k[x]-module via
f (x) · v := f (T )(v). In fact, a k[x]-module is nothing but a vector space with
a specific linear transformation T : V → V .
1.3. Lemma: Let M be an R-module, then for all r ∈ R, m ∈ M ,
(i) 0R · m = 0M
(ii) r · 0M = 0M
(iii) (−r) · m = −(r · m) = r · (−m)
1.4. Definition of submodule N < M
1.5. Examples :

9
 (i) If R is thought of as a module over R, then any submodule is an ideal I C R
(ii) If V is a vector space over k, then any subspace W < V is a k-submodule.
(iii) If G is an abelian group, any subgroup H < G is a Z-submodule.
(iv) If T : V → V is a linear transformation making V into a k[x]-module, then a
k[x]-submodule is a vector subspace W ⊂ V such that T (W ) ⊂ W .
(End of Day 7)

2. Homomorphisms and Quotient Modules

2.1. Definition
(i) Module homomorphism θ : M → M 0
(ii) Isomorphism of modules (In this case, we write M ∼
= M 0)
2.2. Examples :
(i) The zero map
(ii) If G, G0 are Z-modules (abelian groups), then module homomorphism ↔ group
homomorphism
(iii) If V, W are k-modules (vector spaces), then module homomorphism ↔ linear
transformation
(iv) If R is a ring thought of as an R-module, then a module homomorphism is
not the same as a ring homomorphism. For instance x 7→ 2x from Z to Z is a
module homomorphism, but not a ring homomorphism.
(v) If (V, A) and (W, B) are k[x]-modules (as in Example 1.2.(iv)), then a module
homomorphism is a linear map T : V → W such that T A = BT . In particular,
(V, A) ∼= (W, B) iff A and B are similar.
2.3. Definition: θ : M → M 0 a homomorphism of R-modules
(i) ker(θ) < M
(ii) Im(θ) < M 0
2.4. Lemma: θ is injective iff ker(θ) = {0M }
2.5. Theorem: If N < M , then M/N is an R-module via the map (r, a + N ) 7→ ra + N
2.6. (First Isomorphism theorem): Let θ : M → M 0 be a homomorphism of R-modules,
then
M/ ker(θ) ∼
= Im(θ)
In particular, if θ is surjective, then M/ ker(θ) ∼
= M0
2.7. (Second Isomorphism theorem): Let M be an R-module, and L, K < M , then
(i) L + K = {l + k : l ∈ L, k ∈ K} < M
(ii) L/L ∩ K ∼
= (L + K)/K

10
2.8. (Third Isomorphism theorem): Let M be an R-module and L, K < M such that
K ⊂ L, then
(M/K)/(L/K) ∼ = M/L

2.9. (Correspondence theorem): Let M be an R-module, and N < M , then there is

a one-to-one correspondence between the set of submodules of M/N and the sub-
modules of M that contain N
(End of Day 8)

3. Direct Sums of Modules

3.1. Definition
(i) External Direct Sum
Notation: Rn = ⊕ni=1 R
(ii) Internal Direct Sum
3.2. Theorem : If M1 , M2 , . . . , Mn < M , then M is the internal direct sum of M1 , M2 , . . . Mn
iff
(i) M = M1 + M2 + . . . + Mn
P
(ii) For each 1 ≤ i ≤ n, Mi ∩ j6=i Mj = {0}
3.3. Theorem: If M is the internal direct sum of M1 , M2 , . . . , Mn , then

M∼
= M1 ⊕ M2 ⊕ . . . ⊕ Mn

3.4. Definition: Suppose M is the internal direct sum of M1 , M2 , . . . , Mn

(i) Components of m ∈ M
(ii) Projection map πi : M → Mi ⊂ M
3.5. Remark: Suppose M is the internal direct sum of M1 , M2 , . . . , Mn , then
(i) Mi = Im(πi )
(ii) πi2 := πi ◦ πi = πi
Pn
(iii) i=1 πi = idM

4. Finitely Generated Modules

T
4.1. Lemma: Let {Nλ }λ∈Λ be a collection of R-submodules of M , then λ∈Λ Nλ is an
R-submodule.
4.2. Definition of a submodule hXi generated by a set X
Note: By Lemma 4.1, hXi < M

11
4.3. Lemma: Let M be an R-module and X ⊂ M any set, then
∞ X
[ k
hXi = { ri ni : ri ∈ R, ni ∈ N }
k=1 i=1

4.4. Definition:
(i) Finitely generated (f.g.) module
(ii) Cyclic module
4.5. Examples :
(i) (a) R is a cyclic module over R with generator 1
(b) If I C R, then I is cyclic iff I is principal (There is no special name for a
f.g. ideal)
(c) R/I = h1 + Ii is cyclic.
(ii) (a) A k-vector space V is a finitely generated k-module iff it is finite dimen-
sional.
(b) It is cyclic iff dim(V ) ∈ {0, 1}
(c) c0 is an infinite dimensional vector space.
(End of Day 9)
(iii) A cyclic Z-module is a cyclic abelian group. Any finite abelian group is a
finitely generated Z-module. Z is a finitely generated Z-module. Hence,

Z × Z × . . . × Z × Zn1 × Zn2 × . . . × Znk

is a finitely generated Z-module. We will show that all finitely generated

abelian groups look like this.
(iv) Rn ⊕ R/I1 ⊕ R/I2 ⊕ . . . ⊕ R/Ik is a f.g. R-module We will show that, if R is
a PID, then all f.g. R-modules look like this
(v) In particular, if (V, T ) is the f.g. k[x]-module from Example 2.2.(v), then

(V, T ) ∼
= k[x]n ⊕ k[x]/(g1 (x)) ⊕ k[x]/(g2 (x)) ⊕ . . . ⊕ k[x]/(gk (x))

for some ideals I1 , I2 , . . . , In C k[x]. This allows us to determine when two

linear operators are similar.
(vi) A submodule of a f.g. module may not be f.g. : Let R = Cb (R) be the ring of
continuous, bounded functions f : R → R, and consider R as a module over
itself. Let

I = Cc (R) = {f ∈ R : ∃M > 0 such that f (x) = 0 ∀|x| ≥ M }

Then I < R, but I is not f.g., even though R is f.g. (in fact, cyclic)

12
 4.6. Definition : Rank µ(M ) of a module M . If M is not f.g, then µ(M ) = +∞
4.7. Proposition: Let M be an R-module and N < M .
(i) It may happen that µ(M ) < ∞, N < M , but µ(N ) = +∞
(ii) If N < M , then µ(M/N ) ≤ µ(M )
(iii) Suppose N and M/N are f.g., then so if M and µ(M ) ≤ µ(N ) + µ(M/N )
4.8. Definition : Let M be an R-module
(i) Linearly dependent set {x1 , x2 , . . . , xn } ⊂ M
(ii) Linearly independent set
(iii) Torsion-free element
Note: Every non-zero element of a vector space is torsion-free. However, no element
of Zn is torsion-free (when Zn is thought of as Z-module)
4.9. Theorem : Let M be an R-module and X := {x1 , x2 , . . . , xn } ⊂ M be a finite set.
Then, TFAE :
(i) X is linearly independent and M = hXi
P
(ii) For every m ∈ M, ∃!ri ∈ R such that m = ri xi
(iii) Each xi is torsion-free and M ∼
= hx1 i ⊕ hx2 i ⊕ . . . ⊕ hxn i
(End of Day 10)
4.10. Definition :
(i) X generates M freely
(ii) Basis
4.11. Examples:
(i) R = h1i and Rn are freely generated R-modules.
(ii) Any f.g. k-vector space is freely generated by its basis.
(iii) Zn is free as a Zn -module, but not as a Z-module
(iv) Let (V, T ) be a k[x]-module as in Example 1.2.(iv), then (V, T ) is not a free
module since every v ∈ V has torsion.
(v) Every generating set of a vector space contains a basis. However, {2, 3} ⊂ Z
generates Z as a Z-module, but it does not contain a basis.
4.12. Definition :
(i) Annihilator Ann(m). Note: Ann(m) C R
(ii) Torsion module/Torsion-free module
(iii) Mτ is the torsion submodule
4.13. Examples:
(i) A k-vector space is torsion-free as a k-module

13
 (ii) If R is an integral domain, then every non-zero element of R (as a module over
itself) is torsion-free
(iii) (V, T ) is a torsion k[x]-module
(iv) For an Abelian group G, Gτ is the set of elements of finite order. Zn , Q are
torsion-free, and G = Zn are torsion Z-modules; and Z ⊕ Z2 is neither torsion
nor torsion-free.
4.14. Proposition: Let R be an integral domain, and M an R-module. Then
(i) Mτ < M
(ii) M/Mτ is torsion-free

III. Finitely Generated Modules over a PID

Let R be a PID, and M a f.g. module over R

1. Free Modules over a PID

1.1. Definition: Free-rank of M , f (M )
Note: µ(M ) ≤ f (M ), but they may not be equal. Will show that they are equal
when R is a PID.
(End of Day 11)
1.2. Theorem: Let R be a PID, and M a f.g., free R-module. If N < M , then N is a
free R-module and f (N ) ≤ f (M )
1.3. Lemma: Let M be a f.g. R-module, then ∃ a free R-module F such that f (F ) =
µ(M ) and M ∼
= F/N for some submodule N < F
1.4. Corollary: Let M be a f.g. module over a PID R and N < M , then N is f.g. and
µ(N ) ≤ µ(M )
1.5. Lemma: If M is free, then M is torsion-free. (Proof as HW)
1.6. Theorem: A finitely generated torsion-free module is free, and f (M ) = µ(M )
(End of Day 12)
1.7. Theorem: Let M be a free R-module, then
(i) ∃n ∈ N ∪ {0} such that M ∼
= Rn
(ii) Rn ∼
= Rm iff n = m
1.8. Lemma: If θ : M → F is a surjective map where M is f.g. and F is free, then
M∼= F ⊕ ker(θ)
1.9. Theorem: Let M be a f.g. module over R. Then
(i) M/Mτ is free
(ii) ∃n ∈ N ∪ {0} such that M ∼
= Rn ⊕ Mτ

14
 Moreover, n is uniquely determined by M and is called the rank of M .
1.10. Remark:
(i) If n = µ(M/Mτ ), then any set of (n + 1) elements of M is linearly dependent.
(ii) Let M, N be two f.g. R-modules, then M ∼ = N iff

Mτ ∼
= Nτ and rank(M ) = rank(N )

(iii) A torsion module is nothing but a module of rank 0. It now suffices to classify
torsion modules.

2. Torsion Modules over a PID - I

(See [Lang, §III.7]) Let M be a f.g. torsion module over a PID R

2.1. Definition:
(i) For x ∈ M , recall that

Ann(x) := {a ∈ R : ax = 0} C R

Hence, ∃d ∈ R such that Ann(x) = (d). This d is unique upto multiplication

by a unit (by Lemma I.5.2), and is called the order of x, denoted by O(x)
(ii) Recall that
Ann(M ) := {a ∈ R : ax = 0 ∀x ∈ M } C R
By HW 3.3, Ann(M ) 6= {0}, and so ∃a ∈ R such that Ann(M ) = (a). This a
is unique upto multiplication by a unit and is called the exponent of M
(iii) Let p ∈ R prime, then the p-primary component of M is

M (p) := {x ∈ M : pn x = 0 for some n ∈ N}

(End of Day 13)

2.2. Examples :
(i) If G is an abelian group, the order of x ∈ G as a Z-module is the same as the
order of x ∈ G as an abelian group.
(ii) If G is an abelian group, the exponent of G may not be the order of G. For
instance, the exponent of Zp ⊕ Zp is p, but its order is p2
(iii) If G is a finite abelian group, the p-primary component of G is merely the
p-Sylow subgroup of G
(iv) If M = (V, T ) is the k[x]-module, then the exponent of M is the minimal
polynomial of T

15
2.3. Lemma: Let M be a f.g. torsion R-module, with exponent a. Since R is a PID, it
is a UFD (by Theorem I.6.7). Write
a = upα1 1 pα2 2 . . . pαk k
as where u ∈ R∗ and p1 , p2 , . . . , pk are prime elements of R. Then
M = M (p1 ) ⊕ M (p2 ) ⊕ . . . ⊕ M (pk )

2.4. Remark: Suppose M is a f.g. torsion R-module such that M = M (p) for some
prime p ∈ R. Write M = hx1 , x2 , . . . , xn i, then ∃βi ∈ N such that pβi xi = 0 for all
i. Then, for β = max{βi : 1 ≤ i ≤ n}, we have
(i) pβ is the exponent of M
(ii) ∃z ∈ M such that O(z) = pβ . We say that z has maximal order in M
2.5. Lemma: Let M be a f.g. torsion R-module such that M = M (p), and z ∈ M have
maximal order. Write N = hzi, then for any x + N ∈ M/N, ∃y ∈ M such that
O(x) = O(y)
2.6. Theorem: Let M be a f.g. torsion R-module such that M = M (p), then ∃ natural
numbers α1 ≥ α2 ≥ . . . ≥ αk ≥ 1 such that
M∼
= R/(pα1 ) ⊕ R/(pα2 ) ⊕ . . . ⊕ R/(pαk )
(End of Day 14)
2.7. Corollary: Let M be a f.g. torsion module over a PID R, then ∃ prime elements
p1 , p2 , . . . pk ∈ R, and integers {αi,j : 1 ≤ i ≤ k, 1 ≤ j ≤ νi } such that
νi
k M
M∼
α
M
= R/(pi i,j )
i=1 j=1

The list
α α α
{p1 1,1 , p1 1,2 , . . . , p1 1,ν1 ,
α α α
p2 2,1 , p2 2,2 , . . . , p2 2,ν2 ,
.. .. ..
. . .
α α αk,νk
pk k,1 , pk k,2 , . . . , pk }
is called the list of elementary divisors of M , denoted by El(M )
2.8. Corollary: Let M be a f.g. torsion R-module, then ∃d1 , d2 , . . . , dn ∈ R such that
d1 | d2 | . . . | dt
and
M∼
= R/(d1 ) ⊕ R/(d2 ) ⊕ . . . ⊕ R/(dt )
The list
{d1 , d2 , . . . , dt }
is called the list of invariant factors of M

16
3. Torsion Modules over a PID - II
(See [Hartley, §9.2])

3.1. Lemma: For p ∈ R a prime element, and α ≥ 1, let L = R/(pα ). If ϕ, ψ : L → L

be homomorphisms such that
ϕ + ψ = idL
then, either ϕ or ψ is bijective.
(End of Day 15)
3.2. Lemma: If M = M1 ⊕ M2 , and N < M containing M1 , then N = M1 ⊕ (N ∩ M2 )
3.3. (Cancellation Lemma): Let T be a f.g. torsion R-module, and N1 , N2 two R-
modules, then
T ⊕ N1 ∼
= T ⊕ N2 ⇒ N1 ∼ = N2

3.4. (Uniqueness of Invariant Factor decomposition): Suppose d1 , d2 , . . . , dt ∈ R and

d01 , d02 , . . . , d0s ∈ R such that

d1 | d2 | . . . | dt and d01 | d02 | . . . | d0s

and
R/(d1 ) ⊕ R/(d2 ) ⊕ . . . ⊕ R/(dt ) ∼
= R/(d01 ) ⊕ R/(d02 ) ⊕ . . . R/(d0s )
then s = t, and di ∼ d0i for all 1 ≤ i ≤ t.
3.5. Remark: The list of invariant factors is uniquely determined by M , and hence so is
the list of elementary divisors.
3.6. (Structure Theorem for Modules over PIDs) : Let M be a finitely generated module
over a PID R
(i) There exists n ∈ N ∪ {0}, prime elements p1 , p2 , . . . pk ∈ R, and integers
{αi,j : 1 ≤ i ≤ k, 1 ≤ j ≤ νi } such that 1 ≤ αi1 ≤ αi2 ≤ . . . ≤ αiνi and
" k #
M∼
α
M
= Rn ⊕νi R/(p i,j )
j=1 i
i=1

(ii) If N is another finitely generated modules over a PID R, then

(
Elementary Divisors of M = Elementary divisors of N, and
M∼ =N ⇔
rank(M ) = rank(N )

(End of Day 16)

17
4. Finite Abelian Groups
4.1. (Structure theorem for finite abelian groups): Let G be a finite abelian group
(i) There exist prime numbers p1 , p2 , . . . pk ∈ Z, and integers {αi,j : 1 ≤ i ≤ k, 1 ≤
j ≤ νi } such that 1 ≤ αi1 ≤ αi2 ≤ . . . ≤ αiνi and
νi
k M
G∼
M
= Zpαi,j
i
i=1 j=1

(ii) If G0 is another finite abelian group, then

G∼
= G0 ⇔ El(G) = El(G0 )

4.2. Corollary: The number of non-isomorphic abelian groups of order pn , where p ∈ Z

is a prime is equal to π(n), the number of partitions of n.
Example: p = 5, n = 4
4.3. Corollary: Let n = pk11 pk22 . . . pkmm , then the number of non-isomorphic abelian groups
of order n is equal to π(k1 )π(k2 ) . . . π(kn )
Example: Abelian groups of order 600.
4.4. Theorem: Let G be a finite abelian group, and m | |G|, then ∃H < G such that
|H| = m (Proof as HW)
Example: If G = Z8 × Z4 × Z9 × Z3 , then |G| = 864. 144 | |G|, so ∃H < G such
that |H| = 144. We may take H = Z8 × Z2 × Z9 × {0} < G
4.5. Theorem: Let F be a field, and G ⊂ F ∗ = F \ {0} be a finite subgroup of the
multiplicative group F ∗ . Then, G is a cyclic group. In particular, Z∗p is a cyclic for
all primes p
4.6. Example: Let n = p1 p2 . . . pk be square-free, then
(i) There is exactly one abelian group of order n, viz. Zn
(ii) Z∗n ∼
= Zp1 −1 × Zp2 −1 × . . . × Zpk −1
(iii) In particular, if n is divisible by two distinct odd primes, then Z∗n is not cyclic.
In fact, Z∗n is cyclic iff n ∈ {2, 4, pk , 2pk } for some odd prime p (without proof)

5. Rational Canonical Form

(See [Adkins, §4.4])

Let V be a finite dimensional vector space over a field k. Let T : V → V be a linear

transformation. Consider M = (V, T ) as a k[x]-module as in Example II.1.2
5.1. Remark :
(i) By Example II.1.5, There is a 1-1 correspondence between submodules of M
and subspaces of V that are T -invariant.

18
 (ii) By Example II.2.2, There is a 1-1 correspondence between k[x]-module ho-
momorphisms ϕ : (V, T ) → (W, T 0 ) and linear maps S : V → W such that
ST = T 0 S
(iii) (V, T ) ∼
= (V, T 0 ) iff T and T 0 are similar.
(iv) By Example II.4.11, M is a f.g. torsion k[x]-module.
(v) Hence, by Theorem III.3.6, there exist polynomials f1 (x), f2 (x), . . . , fk (x) ∈
k[x] such that f1 (x) | f2 (x) | . . . | fk (x) and

M∼
= k[x]/(f1 (x)) ⊕ k[x]/(f2 (x)) ⊕ . . . ⊕ k[x]/(fk (x))

These fi (x) are unique upto multiplication by a unit. Since k[x]∗ = k ∗ , we

may choose the fi (x) to be monic polynomials, in which case they are simply
unique. The set {f1 (x), f2 (x), . . . , fk (x)} so chosen is called the list of invariant
factors of T
(vi) We define the characteristic polynomial of T to be cT (x) := f1 (x)f2 (x) . . . fk (x)
(vii) We define the minimal polynomial of T to be the unique monic polynomial
mT (x) ∈ k[x] such that Ann(M ) = hmT (x)i. Note: mT (x) = fk (x) and
mT (x) | cT (x)
(End of Day 17)
Notation: Let B be a basis for V , then [T ]B denotes the matrix of T w.r.t. B
5.2. Lemma: Let T : V → V be a linear transformation, and suppose V = V1 ⊕ V2 ⊕
. . . ⊕ Vk , where each Vi is T -invariant. Let Bi be a basis for Vi , then B = ∪Bi is a
basis for V , and  
A1 0 . . . 0
.. 
 0 A2 . 

[T ]B =  .
 ..

0
0 ... Ak
where Ai = [T |Vi ]Bi .
Conversely, if the matrix of T w.r.t some basis B of V has the above form, then V
splits up as the direct sum of T -invariant subspaces as above.
Notation:
(i) T = T1 ⊕ T2 ⊕ . . . ⊕ Tk , where T : V → V is a linear transformation
(ii) A = A1 ⊕ A2 ⊕ . . . ⊕ Ak , where A ∈ Mn (k) is a matrix
5.3. Definition: A linear transformation T : V → V is called cyclic of order f (x) iff
(V, T ) is a cyclic k[x]-module, and exp(V, T ) = f (x)
5.4. Lemma: Suppose T is cyclic of order f (x), write

f (x) = a0 + a1 x + . . . + am−1 xm−1 + xm

then

19
 (i) ∃0 6= v ∈ V such that {v, T (v), T 2 (v), . . . , T m−1 (v)} is a basis for V . In
particular, dim(V ) = deg(f (x))
(ii) The matrix of T w.r.t B is given by
 
0 0 0 . . . 0 −a0
1 0 0 . . . 0 −a1 
 
0 1 0 . . . 0 −a2 
[T ]B =  .. ..
 
.. 
. . . 
 
0 0 . . . 1 0 −am−2 
0 0 . . . 0 1 −am−1

5.5. Definition: Companion matrix C(f ) for a polynomial f (x) ∈ k[x]

5.6. Examples:
(i) C(x − λ) = [λ] ∈ M1 (k)
(ii) diag(a1 , a2 , . . . , an ) = ni=1 C(X − ai )
L
 
2 0 −1
(iii) C(X + 1) =
1 0
(iv) If A = C(X − a) ⊕ C(X 2 − 1), then
 
a 0 0
A = 0 0 1
0 1 0

5.7. Theorem: Let T : V → V , then V has a basis B such that

[T ]B = C(f1 ) ⊕ C(f2 ) ⊕ . . . ⊕ C(fk )
where {f1 (x), f2 (x), . . . , fk (x)} are the invariant factors of T .
(End of Day 18)
5.8. Definition :
(i) Let T : V → V be a linear transformation, then the matrix described in
Theorem 5.9 is called the rational canonical matrix of T
(ii) If A ∈ Mn (k) is any matrix, then define T : k n → k n by T (v) = Av, then the
corresponding matrix for T is called the rational canonical form (RCF) of A.
Note: Any matrix A ∈ Mn (k) is similar to its rational canonical form, and that any
two matrices A, B ∈ Mn (k) are similar iff they have the same rational canonical
form.
5.9. Example: Conjugacy classes in GL2 (Zp )
(i) For A ∈ GL2 (Zp ), write k = Zp , V = Z2p and TA : V → V given by v 7→ Av.
Let M = (V, TA ), then
M∼
= k[x]/(f1 (x)) ⊕ k[x]/(f2 (x)) ⊕ . . . ⊕ k[x]/(fk (x))
Since dim(V ) = 2, it follows that k ≤ 2

20
 (ii) If k = 2, then the RCF of A must belong to
  
λ 0 ∗
C1 = : λ ∈ Zp ,
0 λ

(iii) If k = 1, and f1 (x) is irreducible, then the RCF of A must belong to

  
0 −a0 2
C2 = : a0 + a1 x + x ∈ Zp [x] irreducible
1 −a1

(iv) If k = 1 and f1 (x) is composite with two distinct roots, then the RCF of A
must belong to   
λ 0 ∗
C3 = : λ, µ ∈ Zp , λ 6= µ
0 µ

(v) If k = 1 and f1 (x) is composite with the same repeated root, then the RCF of
A must belong to   
0 −λ2 ∗
C4 = : λ ∈ Zp
1 2λ

(vi) Hence, the representatives for the conjugacy classes in GL2 (Zp ) are

C1 ∪ C2 ∪ C3 ∪ C4

and so the number of conjugacy classes is (p2 − 1) (HW)

(End of Day 19)

6. Cayley-Hamilton Theorem
6.1. Definition:
(i) Characteristic polynomial cT (x) for T : V → V
(ii) Characteristic polynomial cA (x) for A ∈ Mn (k)
6.2. Lemma: Let f (x) ∈ k[x] be a monic polynomial, and A = C(f ) be its companion
matrix, then
det(xI − A) = f (x)

6.3. Lemma: Let A, B ∈ Mn (k) be similar matrices, then cA (x) = cB (x). In particular,
if (V, T ) as before, and B and B 0 are two different bases for V , then

c[T ]B (x) = c[T ]B0 (x)

6.4. Proposition: Let T : V → V be a linear transformation on a finite dimensional

k-vector space as before, then for any basis B of V , we have

cT (x) = c[T ]B (x)

21
6.5. (Cayley-Hamilton Theorem): Let T : V → V be a linear transformation on a finite
dimensional vector space V , and let B be any basis for V . If A = [T ]B , then

cA (T ) = 0

6.6. Remark: Cayley-Hamilton theorem is usually phrased as A linear transformation

satisfies its own characteristic polynomial. Here, the fact that cT (T ) = 0 is trivial
(since mT (x) | cT (x)), but the fact that cA (T ) = 0 is non-trivial.

7. Jordan Canonical Form

Note: If T (x) = λx for all x ∈ V , then
(i) mT (x) = (x − λ) and cT (x) = (x − λ)n
(ii) The rational canonical matrix of T is not a diagonal matrix, and hence not
very useful.
In this section, we assume that k = C
7.1. (Fundamental Theorem of Algebra): The primes in C[x] are of the form (x − λ) for
some λ ∈ C (without proof)
7.2. Remark: Let M = (V, T ) be a C[x]-module as before. The elementary divisor
decomposition of M in Theorem III.2.7 is of the form
νi
k M
M∼
M
= C[x]/(x − λi )nij
i=1 j=1

where {λ1 , λ2 , . . . , λk } are distinct complex numbers, and, for each 1 ≤ i ≤ k, one
has 1 ≤ ni1 ≤ ni2 ≤ . . . ≤ niνi . Consider the subspaces

Vi,j ↔ C[x]/(x − λi )nij

Note that T (Vij ) ⊂ Vij . As in Section III.5, we will find a basis Bi,j of Vi,j such that

Ai,j := [T |Vi,j ]Bi,j

is in a nice form. Then, with B = ∪i,j Bi,j , by Lemma III.5.2, B is a basis for V ,
and M
[T ]B = Ai,j
i,j

7.3. Definition: For λ ∈ C, the n × n Jordan block with value λ is denoted by Jλ,n
(End of Day 20)
7.4. Lemma:
(i) For 1 ≤ k ≤ (n − 1), (Jλ,n − λIn )k 6= 0, but

(Jλ,n − λIn )n = 0

22
 (ii) If V = Cn and T : V → V is given by v 7→ Jλ,n v, then
mT (x) = (x − λ)n = cT (x)

(iii) M = (V, T ) is a cyclic C[x]-module

7.5. Proposition: Let (V, T ) be as before, and suppose M = (V, T ) is a cyclic C[x]-
module with
mT (x) = (x − λ)n
for some λ ∈ C and n ∈ N. Then,
(i) ∃0 6= v ∈ V such that
B = {v, (T − λ)(v), (T − λ)2 (v), . . . , (T − λ)n−1 (v)}
forms a basis for V
(ii) [T ]B = Jλ,n
7.6. Theorem: Let V be a finite dimensional complex vector space and T : V → V be a
linear operator. Then ∃ a basis B of V such that
νi
k M
M
[T ]B = Jλi ,ni,j
i=1 j=1

where {λ1 , λ2 , . . . , λk } are distinct complex numbers and, for each ≤ i ≤ k, 1 ≤

ni1 ≤ ni2 ≤ . . . ≤ niνi .
7.7. Remark: Theorem 7.6 is true over any field k, provided all the irreducible factors
of mT (x) are linear (See HW 5.6)
7.8. Definition:
(i) Jordan Canonical matrix of a transformation T : V → V
(ii) Jordan canonical form of a matrix A ∈ Mn (C).
Note: Two matrices have the same Jordan Canonical form iff they are similar.
(iii) Generalized Eigen-value, Generalized Eigen-vector
(iv) Generalized Eigen-space = (x − λ)-primary component
7.9. Example: Let T : C2 → C2 be a linear transformation, and let cT (x) = (x − λ1 )(x −
λ2 ), then the JCF of T is one of
 
λ1 0
(i) If λ1 6= λ2 :
0 λ2
(ii) If λ1 = λ2 :
 
λ1 0
(a) If mT (x) = (x − λ1 ) :
0 λ1
 
2 λ1 1
(b) If mT (x) = (x − λ1 ) :
0 λ1
(End of Day 21)

23
IV. Introduction to Commutative Algebra
1. Hom and Direct Sums
1.1. Definition: HomR (M, N )
1.2. Examples :
(i) HomR (R, M ) ∼
=M
(ii) HomZ (Zn , Z) = {0} for any n 6= 0
(iii) If M is a torsion R-module, then HomR (M, R) = {0} (HW 6)
(iv) HomZ (Zm , Zn ) ∼= Zd where d = (m, n). In particular, if (m, n) = 1, then
HomZ (Zm , Zn ) = {0}
(v) If V, W are k-vector spaces, then Homk (V, W ) is the collection of linear trans-
formations.
(vi) If (V, T ) and (W, S) are k[x]-modules as before, then by Example II.2.2

Homk[x] ((V, T ), (W, S)) = {U ∈ Homk (V, W ) : U T = SU }

1.3. Definition : Given θ : N → N 0

(i) θ∗ : Hom(M, N ) → Hom(M, N 0 )
(ii) θ∗ : Hom(N 0 , M ) → Hom(N, M )
1.4. Remark/Example :
(i) Pictorial description of Definition 1.3
(ii) (θ ◦ ϕ)∗ = θ∗ ◦ ϕ∗ and (ψ ◦ φ)∗ = φ∗ ◦ ψ ∗
(iii) Consider π : Z → Zn to be the natural projection.

N = Z ⇒ π ∗ = 0 since HomZ (Zn , Z) = {0}

N = Zn ⇒ π ∗ = idZn : Zn → Zn (HW 6)

(End of Day 22)

1.5. Lemma:
(i) Hom(M1 ⊕ M2 , N ) ∼
= Hom(M1 , N ) ⊕ Hom(M2 , N )
(ii) Hom(M, N1 ⊕ N2 ) ∼
= Hom(M, N1 ) ⊕ Hom(M, N2 )
1.6. Corollary: If M and N are free R-modules with m = rank(M ), n = rank(N ), then
Hom(M, N ) ∼ = Rmn
1.7. Definition :
(i) EndR (M )
(ii) AutR (M )
1.8. Examples: AutR (R/I) ∼
= (R/I)∗

24
2. Exact Sequences
2.1. Definition:
(i) A pair of homomorphisms θ : A → B and ψ : B → C forms a chain complex
if ψ ◦ θ = 0; or equivalently, Im(θ) ⊂ ker(ψ)
(ii) A pair of homomorphisms θ : A → B and ψ : B → C is called exact at B if
θ ψ
Im(θ) = ker(ψ). We denote this by A →
− B−
→C
(iii) An exact sequence . . . → Xn−1 → Xn → Xn+1 → . . .
2.2. Lemma:
θ
(i) 0 → A →
− B is exact at A iff θ is injective
ψ
(ii) A −
→ B → 0 is exact at B iff ψ is surjective
θ ψ
2.3. Corollary: 0 → A →
− B−
→ C → 0 is an exact sequence iff
(i) θ is injective
(ii) ψ is surjective
(iii) Im(θ) = ker(ψ)
Such a sequence is called a short exact sequence.
2.4. Example:
(i) Let θ : M → N be any homomorphism, then we have an exact sequence
θ
0 → ker(θ) → M →
− N → N/Im(θ) → 0

(ii) If M = M1 ⊕ M2 , we get a short exact sequence

0 → M1 → M → M2 → 0

(iii) If 0 → V1 → V → V2 → 0 is an exact sequence of vector spaces, then

V ∼= V1 ⊕ V2
(iv) Let p 6= q ∈ Z be prime numbers
(a) 0 → Zp → Zpq → Zq → 0
(b) 0 → Zp → Zp2 → Zp → 0
Both these sequences are exact, but

Zpq ∼
= Zq ⊕ Zq but Zp2  Zp ⊕ Zp

(End of Day 23)

2.5. Definition: A submodule N < M is called a direct summand of M if ∃N 0 < M
such that N ⊕ N 0 = M
θ ψ
2.6. Theorem: For a short exact sequence 0 → M1 →
− M−
→ M2 → 0, TFAE :

25
 (i) There exists a homomorphism α : M → M1 such that α ◦ θ = 1M1
(ii) There exists a homomorphism β : M2 → M such that ψ ◦ β = 1M2
(iii) Im(θ) = ker(ψ) is a direct summand of M
If these conditions hold, then
M∼
= M1 ⊕ M2
via the map x 7→ (α(x), ψ(x)).
2.7. Definition: If the conditions of Theorem 2.6 hold, then we say that the short exact
sequence 0 → M1 → M → M2 → 0 is a split exact sequence.
θ ψ
2.8. Remark: If 0 → M1 →− M−
→ M2 → 0 is an exact sequence of R-modules, and N is
any R-module, then
∗ θ ψ∗
HomR (N, M1 ) −
→ HomR (N, M ) −→ HomR (N, M2 ) (0.1)
is a chain complex. Also,
θ∗ ψ∗
HomR (M1 , N ) ←
− HomR (M, N ) ←− HomR (M2 , N ) (0.2)
is a chain complex. We ask whether these sequences are exact.
θ ψ
2.9. Theorem: For a short exact sequence 0 → M1 →
− M−
→ M2 → 0
In 0.1: (a) θ∗ is injective
(b) ker(ψ∗ ) = Im(θ∗ )
(c) ψ∗ may not be surjective
Hence,
∗ θ ψ∗
0 → HomR (N, M1 ) −
→ HomR (N, M ) −→ HomR (N, M2 )
is exact.
In 0.2: (a) ψ ∗ is injective
(b) ker(θ∗ ) = Im(ψ ∗ )
(c) θ∗ may not be surjective
Hence,
θ∗ ψ∗
HomR (M1 , N ) ←
− HomR (M, N ) ←− HomR (M2 , N ) ← 0
is exact
(End of Day 24)
θ ψ
2.10. Theorem: If 0 → M1 →
− M −
→ M2 → 0 is a split exact sequence, then the induced
sequences
∗ θ ψ∗
0 → HomR (N, M1 ) −
→ HomR (N, M ) −→ HomR (N, M2 ) → 0
and
θ∗ ψ∗
0 ← HomR (M1 , N ) ←
− HomR (M, N ) ←− HomR (M2 , N ) ← 0
are both split exact.

26
3. Projective Modules
(See [Adkins, §3.5]) Assume all modules are f.g.

3.1. Lemma: Let F be a free R-module, then every short exact sequence

0 → M1 → M → F → 0

is split exact.
3.2. Theorem: Let P be an R-module, then TFAE :
(i) Every short exact sequence

0 → M1 → M → P → 0

is split exact
(ii) ∃ a free R-module F s.t. P is a direct summand of F
(End of Day 25)
3.3. Definition: Projective module
3.4. Examples :
(i) Free R-modules are projective
(ii) If (m, n) = 1, then Zm is a projective Zmn -module, but it is not free.
(iii) If P1 , P2 are projective, then P1 ⊕ P2 is projective
(iv) Projective module is torsion free. Hence, a projective module over a PID must
be free (by III.1.6)
√ √
(v) R = Z[ −5] and I = h2, 1 + −5i, then I is projective, but not free.
3.5. Definition: Invertible ideal
3.6. Theorem: Let R be an integral domain, and I C R. Then I is a projective R-module
iff I is invertible.
(End of Day 26)
3.7. Examples:
(i) Every non-zero principal ideal is invertible
√ √
(ii) If R = Z[ −5] and I = h2, 1 + −5i, then I is invertible.
3.8. Remark: A Dedekind domain is an integral
√ domain in which every ideal is invertible.
All PIDs are Dedekind domains, and Z[ −5] is Dedekind domains that is not even
a UFD. Dedekind domains are important in number theory.

27
4. Noetherian Rings
4.1. Definition: R is Noetherian if every ideal is f.g.
4.2. Examples:
(i) Fields, Division Rings, Finite rings
(ii) PIDs
(iii) Cb (R) is not Noetherian (Example I.4.5)
4.3. Lemma: R is Noetherian iff every increasing sequence I1 ⊂ I2 ⊂ . . . of ideals is
stationary. ie. ∃k ∈ N such that In = Ik for all n ≥ k
4.4. (Hilbert Basis Theorem): If R is Noetherian, then R[x] is a Noetherian ring.
4.5. Corollary: If k is a field, then the polynomial ring k[x1 , x2 , . . . , xn ] in n variables is
Noetherian.
4.6. Example: If V = {(0, 0), (0, 1), (1, 0)} ⊂ C2 , and I C C[x, y] is the ideal

I = {f (x, y) ∈ C[x, y] : f (p, q) = 0 ∀(p, q) ∈ V }

Then I = hx2 − x, xy, y 2 − yi

(End of Day 27)
4.7. Definition:
(i) If S ⊂ k n , write I(S) := {f ∈ k[x1 , x2 , . . . , xn ] : f (a) = 0 ∀a ∈ S}
Example: If S = {(0, 0), (0, 1), (1, 0)}, then I(S) = hx2 − x, xy, y 2 − yi
(ii) If I C k[x1 , x2 , . . . , xn ], write V (I) = {a ∈ k n : f (a) = 0 ∀f ∈ I}.
Example: I = hx2 − yi, then V (I) = {(x, y) : y = x2 }
(iii) Hypersurface is V (hf i) =: V (f ) for some f ∈ k[x1 , x2 , . . . , xn ]
4.8. Remark:
(i) Every algebraic set is the intersection of finitely many hypersurfaces
(ii) I(S) C k[x1 , x2 , . . . , xn ]
(iii) I ⊂ I(V (I))
(iv) For 0 6= a ∈ C, and I = hy 2 − x, x − ai, we have I = I(V (I))
(v) If 0 = a ∈ C, and I = hy 2 − x, xi, we have I 6= I(V (I)) = hx, yi. Note that
y ∈ J \ I, but y 2 ∈ I
4.9. (Hilbert Nullstellensatz): If I C C[x1 , x2 , . . . , xn ], then

I(V (I)) = {f ∈ k[x1 , x2 , . . . , xn ] : ∃n ∈ N such that f n ∈ I}

(without proof)
Note:
√
(i) This is called the radical of I, and is denoted by I

28
 √
(ii) If I = {0}, then 0 is the set if nilpotent elements.
4.10. Lemma: If R is Noetherian, and I C R, then R/I is Noetherian.
√ √
4.11. Example: Z[ −5] is Noetherian. In particular, as in HW 2.4, Z[ −5] satisfies
(UF1) (Every non-zero non-unit is a product of irreducible elements)

(End of Day 28)

5. Noetherian Modules
(See [Musili, §6.2])

5.1. Definition: A module M is called Noetherian if every submodule of M is f.g. (In

particular, M is f.g.)
5.2. Examples:
(i) A ring is Noetherian iff it is Noetherian as a module over itself.
In particular, Cb (R) is not Noetherian as a Cb (R) module (See Example I.4.5)
(ii) R[x] is not Noetherian as an R-module
(iii) A module M over a PID if Noetherian iff M is f.g. (See III.1.4)
In particular, a finitely generated abelian group is Noetherian.
A vector space is Noetherian as a k-module iff it is finite dimensional.
(iv) Q/Z is not Noetherian as a Z-module since it is not f.g. (Mid-Sem Exam #
3)
5.3. Prop: Let M be Noetherian, and N < M , then
(i) N is Noetherian
(ii) M/N is Noetherian
5.4. Prop: If M is a module and N < M is s.t. N and M/N are Noetherian, then M is
Noetherian.
5.5. Theorem:
(i) If M1 , M2 , . . . , Mk are Noetherian, then M1 ⊕ M2 ⊕ . . . ⊕ Mk is Noetherian.
(ii) In particular, if R is Noetherian, then Rn is Noetherian
5.6. Corollary: If R is Noetherian, a module over R is Noetherian iff it is f.g.
5.7. Theorem: M is Noetherian R-module iff every increasing chain of submodules is
stationary. (See Lemma 4.3)

29
6. Artinian Modules
(See [Musili, §6.1])
6.1. Definition: Artinian module
6.2. Examples:
(i) If V is a vector space, then V is Artinian iff dim(V ) < ∞
(ii) A finite abelian group is Artinian
(iii) Z is not Artinian.
(iv) Z(p∞ ) is Artinian, but not Noetherian (See HW 5.9)
(End of Day 29)
6.3. Prop: Submodules and Quotients of Artinian modules are Artinian.
6.4. Prop: If M is a module and N < M is a submodule s.t. N and M/N are Artinian,
then M is Artinian.
6.5. Theorem:
(i) If M1 , M2 , . . . , Mk are Artinian, then M1 ⊕ M2 ⊕ . . . ⊕ Mk is Artinian.
(ii) In particular, if R is Artinian, then Rn is Artinian.
6.6. Corollary:
(i) If R is Artinian and M is a f.g. R-module, then M is Artinian.
(ii) An Artinian module need not be f.g. though (Example : Z(p∞ ))
6.7. Theorem: Let R be an Artinian ring
(i) If R is an integral domain, then R is a field.
(ii) Every prime ideal of R is a maximal ideal.

7. Length of a module
(See [Musili, §6.3])
7.1. Definition:
(i) Simple module
(ii) Maximal submodule
(iii) Minimal submodule
7.2. Remark/Examples:
(i) M is simple iff M ∼
= R/I for some maximal ideal I C R (with proof)
(ii) Let V be a k-vector space, then
(a) V is simple iff V is one-dimensional
(b) A subspace W ⊂ V is maximal iff dim(V /W ) = 1. If V is finite dimen-
sional, then this is the same as saying that dim(W ) = dim(V ) − 1

30
 (c) A subspace W ⊂ V is minimal iff dim(W ) = 1
(iii) (a) Zp is a simple Z-module. In fact, this is the only simple Z-module
(b) pZ C Z are the only maximal submodules of Z
(c) Z has no minimal submodules, since for any n ∈ Z, (n) ⊃ (2n)
(iv) Let M = Z(p∞ ), then
(a) M is not simple
(b) M has no maximal submodules
(c) C1 = h1/pi is a minimal submodule
(End of Day 30)
7.3. Lemma: Let M be an R-module
(i) If M is Noetherian, then M has a maximal submodule
(ii) If M is Artinian, then M has a minimal submodule
7.4. Definition: Composition Series
Note: If M has a composition series, then M has both a maximal and a minimal
submodule.
7.5. Examples:
(i) A k-vector space V has a composition series iff dim(V ) < ∞ (with proof)
(ii) A finite abelian group has a composition series.
(iii) Z does not have a composition series
(iv) Z(p∞ ) does not have a composition series, even though it has a minimal sub-
module.
7.6. Theorem: M has a composition series iff M is both Artinian and Noetherian.
7.7. Corollary: Let M be a module and N < M . Then M has a composition series iff
N and M/N both have composition series.
(End of Day 31)
7.8. (Jordan-Hölder Theorem): If M has two composition series

M = M0 ⊃ M1 ⊃ M2 ⊃ . . . ⊃ Mm = {0} and

M = N0 ⊃ N1 ⊃ N2 ⊃ . . . ⊃ Nn = {0}
then
(i) n = m
(ii) For all 1 ≤ i < m, ∃1 ≤ j < n such that Mi /Mi+1 ∼
= Nj /Nj+1
7.9. Definition: Length of a module M , denoted by `(M )
7.10. Remark/Examples:

31
 (i) `(M ) ≥ 0 and `(M ) = 0 iff M = {0}
(ii) `R (M ) = 1 iff M is simple. In general, `(M ) is the departure of M from being
simple.
(iii) If G is a finite abelian group and |G| = p1 p2 . . . pk (where the pi are not
necessarily distinct primes), then `(G) = k
(iv) If V is a k-vector space, then `(V ) = dim(V )
7.11. Theorem: Let M be an R-module and N < M . Then `(M ) = `(N ) + `(M/N )
7.12. Corollary: If M := M1 ⊕ M2 ⊕ . . . ⊕ Mk , then `(M ) = ni=1 `(Mi )
P

7.13. Corollary: Let V be a k-vector space, then TFAE :

(i) V is finite dimensional
(ii) V has finite length
(iii) V is Noetherian as a k-module
(iv) V is Artinian as a k-module
If any of these conditions are satisfied, then any two bases of V have the same
number of elements, and `(V ) = dim(V )

(End of Day 32)

V. Tensor Products
1. Finite Dimensional Vector Spaces
(See [Gowers])

Let U, V, W, X, etc. denote finite dimensional vector spaces over a field k

1.1. Definition: Bilinear map f : V × W → X
1.2. Examples:
(i) If V is an inner product space over R, then the inner product h·, ·i : V ×V → R
is bilinear.
(ii) Cross product R3 × R3 → R3
(iii) If V is a vector space, and V ∗ its dual, then B : V × V ∗ → k defined by
B(v, f ) := f (v) is bilinear.
(iv) ψ : C × Rn → Cn given by (z, v) 7→ (zv1 , zv2 , . . . , zvn )
1.3. Remark:
(i) A linear map T : V → W can be encoded in a matrix [T ]B for some basis B of
V . Similarly, a bilinear map f : V ×W → X can be encoded in a 3-dimensional
n × m × r grid, whose (i, j, k)th entry is the k th coordinate of f (vi , wj ) where
{vi } and {wj } are bases of V and W respectively.

32
 Question: Can we instead encode f in a matrix? In other words, we need a
suitable set S of vectors such that [f ]S captures f uniquely.
(ii) If V = W = X = R, then S = {(1, 0), (0, 1)} forms a basis for V × W ,
but [f ]S is zero, and does not give any information regarding f . However, if
0 6= a, 0 6= b, then S = {(a, b)} is not a basis for V × W , but f (a, b) gives all
the information we need about f since f (x, y) = xy ab
f (a, b) for any bilinear f
(iii) Any f : V ×W → X is determined by f = (f1 , f2 , . . . , fr ) where fi : V ×W → k
are all bilinear.
1.4. Definition:
(i) Bk (V, W ) is the vector space of all bilinear maps f : V × W → k
(ii) For v ∈ V, w ∈ W , define v ⊗ w : Bk (V, W ) → k by v ⊗ w(f ) := f (v, w).
Notice that v ⊗ w ∈ Bk (V, W )∗ , the dual space of Bk (V, W )
(iii) Define V ⊗ W := span{v ⊗ w : v ∈ V, w ∈ W }
1.5. Theorem: If {vi } and {wj } are bases for V and W respectively, then {vi ⊗ wj } is a
basis for V ⊗ W . In particular, dim(V ⊗ W ) = dim(V ) × dim(W )
1.6. Lemma: The map ϕ : V × W → V ⊗ W given by ϕ(v, w) := v ⊗ w is bilinear.
1.7. Proposition: If X is a finite dimensional vector space, and g : V × W → X is a
bilinear map, then ∃!T : V ⊗ W → X linear such that T ◦ ϕ = g. In other words,
there is an isomorphism

BX (V, W ) ∼
= Homk (V ⊗ W, X)

(End of Day 33)

1.8. Theorem: The pair (V ⊗ W, ϕ) is unique in the following sense : If U is a finite
dimensional vector space and ψ : V × W → U is a bilinear map such that, for any
bilinear map h : V × W → X, ∃!S : U → X such that S ◦ ψ = h, then there is an
isomorphism µ : U → V ⊗ W such that µ ◦ ψ = ϕ
1.9. Example:
(i) C ⊗ Rn ∼= Cn
(ii) V ⊗ V ∗ ∼
= Endk (V )
(iii) If ψ : V × V ∗ → k is given by (v, f ) 7→ f (v), then the induced linear map
S : Endk (V ) → k is the trace.

2. Tensor Product of Modules

Note: If M and N are modules, then M ⊗ N cannot follow the same idea as above,
because if M has torsion, then BR (M, N ) = {0}
2.1. Definition for modules (in terms of universal property)
2.2. Theorem: The tensor product (T, ϕ), if it exists, is unique.
2.3. Examples:

33
 (i) R ⊗R M ∼=M
(ii) Za ⊗Z Zb ∼
= Zd where d = (a, b)
In particular, if (a, b) = 1, then Za ⊗Z Zb = {0}
(End of Day 34)
2.4. Remark: (The construction of M ⊗R N )
(i) Let F be the free module on M × N
(ii) Let H be the subgroup generated by

(m1 + m2 , n) − (m1 , n) − (m2 , n)

(m, n1 + n2 ) − (m, n1 ) − (m, n2 )

(rm, n) − r(m, n)
(m, rn) − r(m, n)

(iii) Define T := F/H and ϕ : M × N → T by (m, n) 7→ (m, n) + H

2.5. Theorem: The pair (T, ϕ) is a tensor product of M and N . We write T = M ⊗R N
and m ⊗ n := ϕ(m, n)
2.6. Theorem:
(M1 ⊕ M2 ) ⊗ N ∼
= (M1 ⊗ N ) ⊗ (M2 ⊗ N )
M ⊗ (N1 ⊗ N2 ) ∼
= (M ⊗ N1 ) ⊕ (M ⊗ N2 )

2.7. Examples:
(i) Zn ⊗Z Q = {0}
So, If G is any finite abelian group, then G ⊗ Q = {0}
(ii) Zn ⊗ Q ∼= Qn (Similar to Example 1.9(ii))
If G is a finitely generated abelian group, then G⊗Q is a Q-vector space whose
dimension is the free-rank of G
(iii) More generally, if M is a f.g. module over a PID R, and K = Quot(R), then
M ⊗ K is a K-vector space whose dimension os the free-rank of M
2.8. Corollary: If M and N are free R-modules of rank m and n respectively, then
M ⊗ N is a free module of rank mn
(End of Day 35)
2.9. Remark: (Extension of Scalars)
(i) If R ⊂ S are two rings, and M is an S-module, then M is also an R-module.
But if M is an R-module, then can M be made into an S-module?
(ii) Example:
(a) Z is a Z-module, but cannot be made into a Q-module: If it could, then
x = 21 · 1 ∈ Z would satisfy 2x = 1 which is impossible. However, Z is
contained in a Q-module, namely Q itself.

34
 (b) If M = Z/2Z is a Z-module, then M does not embed in any Q-module,
because if N is any Q-module and f : M → N any embedding, then
2f (x) = 0 for all x ∈ M , and hence 21 · 2f (x) = 0 ⇒ f ≡ 0
(iii) The map S × S ⊗ M → S ⊗ M given by
n
X n
X
(s, si ⊗ mi ) 7→ (ssi ) ⊗ m
i=1 i=1

makes S ⊗ M into an S-module.

2.10. Theorem: Let ι : M → S ⊗ M be the map m 7→ 1S ⊗ m. If N is any S-module
(and hence R-module), and θ : M → N a homomorphism of R-modules, then
∃!T : S ⊗ M → N such that T ◦ ι = θ. Hence,
HomR (M, N ) ∼
= HomS (S ⊗ M, N )

Review of Chapter I
(End of Day 36)
Review of Chapter II, III, IV
(End of Day 37)
Review of Chapter IV, V
(End of Day 38)

VI. Instructor Notes

0.1. Doing Chinese Remainder theorem in this course (as against MTH 301) was hugely
beneficial. It supports the structure theorem for modules, and helps understand
the transition between invariant factors and elementary divisors. Doing it in MTH
301 would not have been as helpful.
0.2. I discussed Rational/Jordan canonical forms as an important application of the
structure theorem because it hadn’t been covered in Advanced Linear Algebra, and
also because it served as an important example throughout the course.
0.3. I discussed composition series because I wanted them to see Jordan-Hölder once
before seeing it in Galois theory when discussing solvable groups. We will see next
semester if this was beneficial or not.
0.4. I did not discuss non-commutative rings - Schur’s Lemma,etc - but decided to focus
on commutative algebra in the form of Noetherian/Artinian modules instead. I
believe the course flows better this way, but perhaps if I had had more lectures
(we lost many due to various holidays), I would have been able to discuss non-
commutative rings as well.
0.5. Although I discussed tensor products, it was woefully inadequate. Perhaps it was
my fault, but I think they are a little too abstract for this point in the students’
education. Not sure it should be there in this course.

35
Bibliography
[Norman] Christopher Norman, Finitely Generated Abelian groups and Similarity of Ma-
trices over a Field

[Hartley] Hartley and Hawkes, Rings, Modules and Linear Algebra

[Adkins] Adkins and Weintraub, Algebra : An approach via Module theory

[Dummit] Dummit and Foote, Abstract Algebra

[Lang] S. Lang, Algebra, Revised 3rd Ed.

[Musili] C. Musili, Introduction to Rings and Modules, 2nd Revised Edition

[Gowers] T. Gowers, How to lose your fear of tensor products (Online Notes)

36