Algebra
Uploaded by
mishusaini3012Algebra
Uploaded by
mishusaini30126
MODULES
Structure
6.1. Introduction.
6.2. Module.
6.3. Homomorphism of Modules.
6.4. Minimal Generating Set.
6.5. Torsion Element.
6.6. Check Your Progress.
6.7. Summary.
6.1. Introduction. In this chapter Module theory is discussed in detail. The concepts of generating sets,
rank of a finitely generated module and direct sum of two submodules are detailed.
6.1.1. Objective. The objective of these contents is to provide some important results to the reader like:
(i) Simple Modules, Unital Modules, Free Module, Irreducible Module, Quotient Modules.
(ii) Submodules.
(iii) Kernel and Null Space of a homomorphism.
(iv) Schurs’ Lemma.
(v) Fundamental Theorem on Finitely Generated Modules.
Keywords. Modules, Submodules, Free Module. Rank.
6.2. Module. Let R be a ring. A non empty set M is said to be a left module R (or a left R - module) if
M is an abelian group under an operation ‘+’ such that for every r R , m M , there exists a unique
element rm M subject to the conditions:
(i) r(a+b) = ra+rb
(ii) (r+s)a = ra+sa
(iii) r(sa) =(rs)a
78 Modules
In a similar way, we can define a right R – module by modifying the conditions (i), (ii), (iii) in the above
definition in the following manner,
(i) (a+b)r =ar + br
(ii) a(r + s) =ar + as
(iii) (ar) s = a (rs)
6.2.1. Note. The theory of right R-modules can be developed in same manner as the theory of left R-
module . We shall develop here the theory of left R-modules and shall be omitting the adjective left.
6.2.2. Unital R-Module. If R is a ring with unity, then a R – module is said to be unital if
1.m = m for all m M.
Remark. If a ring R is a field, then a unital R-Module is a vector space over the field R . So we can say
that concept of a module is a generalisation of the concept of a vector space.
6.2.3. Example. Every abelian group G is a module over the ring of integers, Z.
Solution. Let G be an abelian group, the operation in G being denoted by + and identity of G being 0.
For any integer n and for element a ϵ G, we define
na = a + a +…+ a (n-times)
Then by closure property na ϵ G.
Now to prove that G is a module over I, we have to prove that
(i) m( a + b) =( a + b) + ( a + b) +… +( a + b) (m-times)
=(a + a +… + a) + (b + b +… + b) [ G is abelian]
= ma + mb
(ii) (m + n) a = a + a +… + a (m + n times)
= (a + a +… + a) + (a + a +… +a)
m- times n- times
= ma + na
(iii) m(na) = na + na +… +na (m-times)
= n(a + a +… + a) ( m-times)
= ( a + a … + a ) + ( a + a …+ a ) + ...+ ( a + a …+ a) [ n-times]
= a + a +…+ a ( mn - times)
= (mn) a
Hence G is a module over I.
Abstract Algebra 79
6.2.4. Exercise.
1. Every ring R is an R – module over itself.
2. Every ring R is an module over its subring S, OR, If R is a ring and S be its subring then R is an S –
module.
Remark. With this, we confirm that ring R of real numbers is a Q – module and a Z – module.
3. Every abelian group G is module over the ring of integer Z.
4. Let R be a ring and n be a positive integer. Then R n ( a1 , a2 ,..., an ) : ai R is an R – module
under the operations defined by
(a1 , a2 ,..., an ) (b1 , b2 ,..., bn ) (a1 b1 , a2 b2 ,..., an bn )
and r (a1 , a2 ,..., an ) (ra1 , ra2 ,..., ran )
for all (a1 , a2 ,..., an ), (b1 , b2 ,..., bn ) R n and for all r R .
5. (Elementary properties of module) Let R be a ring and M be an R – module. Then for all a, b, c M ,
we have
(i) a b a b 0
(ii) a b 0 a b
(iii) a b a c b c
(iv) b a c a b c
6. Let R be a ring and M be an R – module . Then
(i) r 0 0 for all r R
(ii) 0a 0 for all a M
(iii) ( r ) a (ra ) r ( a ) for all r R , a M
(iv) r (a b) ra rb for all r R , a, b M
(v) ( r s )a ra sa for all r , s R, a M .
7. Let R be any ring and λ be a left ideal of R. Let M consist of all cosets a + λ, where a ∈ R, of λ in R
Thus, M = { a + λ : a ∈ R} Then, M is an R- module if the two requisite composition are defined as
follows:
(a + λ) + (b + λ) = (a + b) + λ
r(a + λ) = ra + λ
Remark. M is written as R- λ or R/λ and is called the difference (or quotient) module of R by λ.
80 Modules
6.2.5. Sub- module. A non – empty subset N of an R – module M is said to be submodule of M if N
itself is an R – module under the operations of addition and scalar multiplication as defined for M.
If M is an R-module then M itself and {0} are always submodules of M and are termed as improper
submodules. Any submodules of M other than M and {0} are called proper submodule.
6.2.6. Exercise.
1. Let R be a ring and let M be an R – module. A non –empty subset N of M is a submodule of M iff N
is additive subgroup of M and is closed under scalar multiplication
Alternatively, A non – empty subset N of R – module M is a submodule of M iff a b N for all
a, b N and ra N for all a N , r R .
2. If A and B are two submodule of an R- module M. Then A∩B is also a sub-module of M.
3. Arbitrary intersection of submodules of a module is a submodule.
4. Let {Mλ} be a family of submodules of module M. Suppose this family is totally ordered or linearly
ordered, that is, given M 1 , M 2 {M } then either M 1 M 2 or M 2 M 1 then M is a
submodule of M.
5. If M is an R – module and a M then the set Ra {ra : r R} is a sub module of M.
6. Let M be an R – module. Define S = {ra ma : r R , m Z} , Z being the ring of integers, then S
is an R – submodule of M containing ‘a’.
7. If A and B are submodules of an R-module M, then A+B is also a submodule of M .
6.2.7. Submodule generated by a subset of module. Let M be an R- module and S be a non-empty
subset of M. If A is a submodule containing S and is itself contained in every submodule of M
containing S, then A is called the submodule of M generated by [Link] submodule of M generated by S
will be denoted by < S >. It should be noted that < S > is the smallest submodule of M containing S. It
can be seen that the intersection of all the submodule of M containing Sis the submodule of M generated
by S.
6.2.8. Theorem. The submodule of a unital R- module M generated by a subset S of M consists of all
linear combinations of elements in S.
Proof. Let L(S) denote the set of all linear combinations of the elements of S, that is,
L(S)= {r1a1 + r2a2 + . . . + rnan : ai ∈ S, ri ∈ R}
First we shall prove that L(S) is a sumodule of M.
Let a = r1a1 + r2a2 + . . . + rnan and b = s1b1 + s2b2 + . . . + snbn be any two elements of L(S), where ri, si
∈R and ai, bi ∈ S.
We have,
a-b = r1a1 + r2a2 + . . . + rnan + (-s1)b1 + (-s2)b2 + . . . + (-sn)bn
is a linear combination of some elements of S, which implies a-b ∈ S. Thus, L(S) is an additive
subgroup of M.
Abstract Algebra 81
If r ∈ R and a = r1a1 + r2a2 + . . . + rnan ∈ L(S), then
ra = r(r1a1 + r2a2 + . . . + rnan)
= r(r1a1) + r(r2a2) + . . . + r(rnan) = rr1(a1) + rr2(a2) + . . . + rrn(an)
which implies, ra ∈ L(S), as rr1, rr2, . . . rrn ∈ R. Hence, L(S) is a submodule of M.
Next, we claim that S is contained in L(S).
Let a ∈ S, then a ∈ S, 1 ∈ R, so a.1 ∈ L(S), or, a ∈ L(S). Hence S is contained in L(S).
Thus, L(S) is a submodule of M containing S.
Now, if W is any submodule of M containing S,then each element of L(S) must be in W, as W is closed
under scalar multiplication & addition.
Therefore, L(S) will be contained in W. Hence, L(S) = < S >, that is, L(S) is the smallest submodule
generated by S.
6.3. Homomorphism of Modules.
Let M and N be two R-modules. A mapping T : M → N is called a homomorphism or module
homomorphism if
(i) T(m1+m2) = T(m1) + T(m2) for all m1,m2 ∈ M
(ii) T(rm) = rT(m) for all m ∈ M, r ∈ R
The kernel K(T) of T is defined as
K(T ) ={m ∈ M: T(m) = 0 where 0 is the additive identity of N }
8.3.1. Theorem. The kernel of a homomorphism is a submodule.
Proof. Let K(T) be the kernel of homomorphism T of R-module M in to an R-module N, then K(T) =
{m ∈ M : T(m) = 0}. We are to prove that K(T) is a submodule of M. Since
T(0) = 0 => 0 ∈ K(T) => K(T) ≠ ∅
Let m1,m2 ∈ K(T) => T(m1) = 0 = T(m2), then T(m1 - m2) = T(m1) - T(m2) = 0 - 0 = 0
=> m1- m2∈ K(T)
Hence, K(T) is an additive subgroup of M.
Again, r ∈ R and m ∈ K(T) => T(m) = 0, then T(rm) = rT(m) => r.0 = 0 => rm ∈ K(T)
Hence, K(T) is a submodule of M.
6.3.2. Theorem. The range of a homomorphism is a submodule.
Proof. Let T: M → N is a homomorphism where M and N are R-submodules. Then, T(M) is the range
of M under T. So, T(M) = {T(m) : m ∈ M}.
82 Modules
We claim that T(M) is a submodule of N. Let T(m1) and T(m2) be any two element of T(M) where
m1,m2 ∈ M
We have, T(m1 - m2) = T(m1) - T(m2) => T(m1 - m2) ∈ T(M) [ m1-m2 ∈ M ]
∴ T(M) is an addtive subgroup of N.
Let r ∈R and T(m) ∈T(M) where m ∈M, then we have rT(m) = T(rm ) ∈ T(M), as rm ∈ M.
Hence T(M) is a submodule of N.
6.3.3. Theorem. Let T:M→N be a module homomorphism. Then, T is a monomorphism iff K(T) = {0}
Proof. Let T be homomorphism of an R-module M in to an R-module N and assume that K(T) = {0}.
We claim that T is a monomorphism
If m1, m2 ∈ M such that T(m1) = T(m2)
=> T(m1) - T(m2) = 0 => T(m1- m2) = 0
=> m1- m2∈ K(T) ={0} => m1- m2 = 0
=> m1= m2 => T is one-one
Conversely, suppose that t is one-one
Let m ∈ K(T) then T(m) = 0. We know that T(0) = 0 => T(m) =T(0) => m =0, since T is one-one.
Hence Ker T = K(T) = {0}.
6.3.4. Quotient Modules. Let A be any submodule of an R-module M. Then A is an additive abelian
subgroup of M. If m ∈ M, then m + A is coset of A in M. Then,
M/A = { m + A : m ∈ M }
is an R-module unknown as quotient module with addition and scalar multiplication defined as:
(m1 + A) + (m2 + A) = (m1+m2) + A
and r(m1 + A) = rm1 + A; r ∈ R, m1 ∈ M.
6.3.5. Exercise.
1. If M is an R-module and N is an R-submodule of M. The mapping T : M → M/N defined as
T(m)=m+N for all m ∈ M. Then, T is an R- homomorphism of M onto M/N and kerT = N.
2. Fundamental Theorem of Homomorphism on Modules. If T is a homomorphism of an R-module
M onto an R-module N such that ker T = A then N is isomorphic to M/A, that is, M ker T N .
Hint. To derive it define a mapping Ψ: M/A →N as Ψ(m + A) = T(m) for all m ∈ M, and then prove that
this mapping is an isomorphism.
6.3.6. Cyclic Module. An R-module M is said to be cyclic if there is an element mo ∈ M such that every
m ∈ M is of the form m = rmo, where r ∈ R . Also mo is called a generator of M and we can write M = <
mo > = {rmo : r ∈ R}.
Abstract Algebra 83
6.3.7. Theorem. Let M be a unital R-module and for a fixed element m ∈ M, let A = {rm : r ∈ R}. Then,
A is cyclic submodule of M generated by m.
Proof. Let a, b ∈ A Then, a = r1m and b= r2m for r1,r2 ∈ R. We have
a - b = r1m - r2m = (r1- r2)m ∈A [ r1- r2∈ R]
For r ∈ R and a ∈A such that a = r1m for some r1∈R
ra = r(r1m)=(rr1)m ∈ A [ rr1 ∈R]
=> ra ∈A
Hence, A is a submodule of M. Also 1 ∈ R, so 1.m ∈ A => m ∈ A. Due to the definition of A, we
conclude that A is cyclic submodule of M generated by m.
6.3.8. Theorem. If A and B are submodules of M. Then, A B A
B A B .
Proof. Consider the mapping Ψ: A+ B → A
A B defined by
Ψ(x + y ) = x + (A B) for all x ∈ A, y ∈ B.
(i) Ψ is well defined. Let x1 + y1, x2 + y2 ∈ A + B are such that
x1 + y1= x2 + y2 => x1 - x2 = y2 - y1
Since A and B are submodules of M, so x1, x2 ∈ A and y1, y2 ∈ B => x1 - x2 ∈ A, y2 - y1 ∈ B.
Now, x1 - x2 = y2 - y1 => x1 - x2 ∈ B [ y2 - y1∈B ]
and y2 - y1 ∈ A [ x1- x2∈A]
=> x1 - x2 = y2 - y1 ∈ A B
=> x1 - x2 ∈ A B
=> x1 + A B = x2 + A B
=> Ψ(x1+ y1) = Ψ( x2 + y2)
(ii) Ψ is a module homomorphism. Let x1 + y1, x2 + y2 ∈ A + B. Then
Ψ[(x1+y1) + (x2+ y2)] = Ψ[(x1 +x2)+(y1 + y2)]
=( x1+ x2) + A B= (x1 + A B) + (x2 + A B)
= Ψ(x1+y1) + Ψ (x2+ y2)
and Ψ[r(x+y)]= Ψ(rx+ry) = rx + A B
= r(x + A B)=r Ψ(x+y)
Therefore, Ψ is a module homomorphism.
84 Modules
(iii) Ψ is onto.
Let x+ A B ∈ A
A B , where x ∈ A => x + y ∈ A+B for y ∈ B and Ψ(x+y ) = x+ A B
=>Ψ is onto.
By Fundamental Theorem of module homomorphism, we have
A B A (*)
ker A B
We claim that ker Ψ = B.
By definition, ker Ψ = {x+y : Ψ(x+y)= A B} = {x+y : x+A B = A B}
= {x+y : x ∈ A B}={x+y : x ∈ B, y ∈ B}
={z : z ∈ B} = B
Hence by (*), we have A B A
B A B .
6.3.9. Irreducible Module. A R-module having no proper submodule is called an Irreducible module.
6.3.10. Theorem. Prove that any unital, Irreducible R-module is cyclic.
Proof. Let M be any unital irreducible R-module, then the only submodule of M are <0> and M itself.
We claim that M is cyclic
If M = < 0 >, then obviously M is cyclic. Let M ≠ < 0 > . Then, there exists atleast one element mo ∈ M
such that m0 ≠ 0 .
Let A ={rmo : mo∈ M, r ∈ R}
We shall prove that A is a submodule of M
Let a, b ∈ A, then a= r1mo and b = r2mo for r1, r2 ∈ R and mo ∈ M.
We have, a - b = r1mo - r2mo = (r1- r2)mo ∈ A => a - b ∈ A
Therefore, A is an additive subgroup of M.
Now let r ∈ R and a= r1mo ∈ A.
Then ra = r(r1mo) = (rr1)mo ∈ A [ rr1 ∈R]
=> ra ∈ A for all r ∈R
=> A is submodule of M
As given M is unital module, so [Link] = mo, where 1 is unity of R .
Since 1 ∈ R, mo ∈ M => [Link] ∈ A => mo ∈ A.
However, mo ≠ 0, so A ≠ < 0 >, but M is irreducible. So we must have A=M.
Since A is cyclic, so M must be cyclic.
Abstract Algebra 85
6.3.11. Exercise.
1. If λ is a left ideal of R and M is an R-module, then for m ∈ M, λm={xm : x ∈λ} is a submodule of M.
2. Let M be an R- module. If I is a right ideal of R, then the collection of elements y ∈ M such that by =
0 for all b ∈ I is a submodule of M.
Hint. Let A = {y : y ∈M, by = 0 for all b ∈ I }, Then A is a submodule of M
6.3.12. Theorem. Suppose that R is a ring with unity and M is a module over R but is not unital. Then,
there exists some non-zero m in M such that rm = 0 for all r ∈ R.
Proof. Since M is not unital then there exists m ∈ M such that 1.m ≠m
=>1.m - m ≠0 But 1.m - m ∈ M [ 1.m∈M as 1 ∈ R, m ∈ M]
Let m0 = 1.m -m ≠ 0
Then rm0 = r(1.m - m) =r(1.m) - rm = 0, that is, rm0 = 0 where m0 ≠ 0 for all r ∈ R
In particular, rm0 = 0 for all r ∈ R.
6.3.13. Exercise. If M is an irreducible R- module. Then, either M is cyclic module or for every m ∈ M
and r ∈ R, rm = 0.
Solution. Given M is irreducible, that is, < 0 > and M are the only submodules of M.
Let m ∈ M. Consider N= {rm : r ∈ R }
We claim that N is a submodule of M
Let α, β ∈ N => α = r1m, β = r2m for some r1, r2 ∈ R. Then α - β = r1m - r2m = (r1 - r2)m ∈ N, as r1 - r2 ∈
R. Therefore, N is an additive subgroup of M.
Let α ∈ N and r ∈ R then rα= r (r1m) =(rr1)m ∈ N. [ rr1 ∈R]
Thus, N is a submodule of M .
=> N = < 0 > or N = M
If N = < 0 > then rm=0 for all r ∈ R, m ∈ M.
If N = M, then M = {rm: r ∈ R, m ∈ M }
=> M is a cyclic submodule.
6.3.13. Exercise.
1. Suppose M and N are submodules of a module P over R, then
M N {0}
iff every element z ∈ M+N can be uniquely expressible as z=x+y with x ∈ M, y ∈ N.
2. The necessary and sufficient condition for a module M to be the direct sum of its two submodules
M1 and M2 are that (i) M = M1 +M2, (ii) M1 M2 = {0}.
86 Modules
Finitely Generated Module. An R-module M is said to be finitely generated if there exist elements a1,
a2, . . ., an M such that every m in M can be written as:
m = r1a1 + r2a2 + … + rnan; ri’s R
Then, {a1, a2, . . ., an} is called the generating set of M.
6.4. Minimal Generating Set. The generating set of a module which has as few elements as possible is
called minimal generating set, that is, if we remove even a single element from this set, this set is no
longer a generating set.
6.4.1. Rank of a Module. The number of elements in a minimal generating set of module is known as
the rank of module.
6.4.2. Fundamental Theorem on Finitely Generated Module.
Let R be a Euclidean ring, then any finitely generated R-module M, is the direct sum of a finite number
of cyclic modules.
Proof. We shall prove the theorem by induction on the rank of M. If rank of M is 1, then M is generated
by a single element and hence M is cyclic. The theorem is true in this case. Now as an induction
hypothesis, we assume that the theorem holds for all R-modules of rank k - 1.
Now consider R-modules M with rank k.
Given any minimal generating set a1, a2, …, ak of M, if any relation of the form
n1a1 = n2a2 = … = nkak = 0,
then M is the direct sum of M1,M2,….,MK where each Mi is cyclic generated by ai and the theorem is
proved.
So, let given any minimal generating set b1, b2, …, bk there exists r1, r2, …, rk R such that r1b1 + r2b2 +
… + rkbk = 0 but not all of r1b1, r2b2, …, rkbk are zero.
Now among all possible such relations for all minimal generating sets, let s1 be the element of R with
minimum d-value d(s1) and let the generating set for which it occurs be a1, a2, …, ak. Thus we have
s1a1 + s2a2 + … + skak = 0. (1)
Now, if for any r1, r2, ..., r ;
r1a1 + r2a2 + ... + rkak = 0, (2)
then we claim that s1/r1.
Since r1, s1 ∈ R, a Euclidean ring, so there exist m, t ∈ R with r1 = ms1 + t, where either t = 0 or d(t) <
d(s1).
Multiplying (1), by m and subtracting from (2), we get
(r1 - ms1)a1 + (r2 - ms2)a2 + ... + (rk - msk)ak = 0 => ta1 + (r2 - ms2)a2 + . . . + (rk - msk)ak = 0
Thus by choice of s1 we must have d(s1) < d(t )
=> d(t) d(s1) => t =0 => r1= ms1 => s1/r1.
Abstract Algebra 87
We further claim that s1/si for i = 2, 3, …, k.
Suppose not, then let s1 does not divide s2, so there exists m2, t2 ∈ R such that s2 = m2s1 + t2, where d(t2)<
d(s1) [if t2 =0 then s1/s2].
Now clearly a1' a1 m2 a2 , a2 ,..., ak generates M.
Consider, s1 a1' + t2a2 + s3a3 + … + s kak = s1(a1 + m2a2) + t2a2 + s3a3 + ... + skak
= s1a1 + (s1m2 + t2)a2 + s3a3 + ... + skak
= s1a1 + s2a2 + s3a3 + ... + skak = 0.
Thus, t2 occurs as a coefficient in some relation among elements of a minimal generating set. so by
choice of s1 we have d(s1) < d(t2), which is a contradiction to the statement that
d(t1) < d(s1)=> s1/s2 .
Similarly it can be proved that s1/si for all i.
Let s2 = m2s1, s3 = m3s1, ..., sk = mks1, [2(a)]
Now consider the elements a1* a1 + m2a2 + ... + mkak, a2, a3, …, ak. Clearly a1* , a2, ..., ak generate M.
Let M1 be the cyclic submodule of M generated by a1* and M2 be the submodule of M generated by a2,
...,ak. We claim that M = M1 M2 (*)
Since a1* , a2, ...,ak generate m, so clearly M = M1 +M2 in order to prove (*) only we have to prove now is
that M1 M2 = {0}. For this, let p ∈ M1 M2
=> p ∈ M1 and p ∈ M2
Now, p ∈ M1 => p = r1 a1* , for some r1 ∈ R [ M1 is generated by a1* ]
Also, p ∈ M2 => p = r2a2 + ... + rkak for some r2, r3, ..., rk ∈ R
Now we have r2 a1* r2 a2 + ... + rkak,
=> r1 a1* - r2 a2 + ... + rkak = 0
=> r1(a1 + a2m2 + ... + akmk) - r2a2 - ... - rkak = 0
=> r1a1 + (r1m2 - r2)a2 + ... + (r1mk - rk)ak = 0 (3)
Thus we have obtained a relation between a1, a2, ..., ak in which coefficient of a1 is r1. Hence by what we
have proved above s1/r1. Let r1 = ls1 where l ∈ R, so we have
p = r1a1*
= l[s1(a1 + m2a2 + ... + mksk)] = l[s1a1 + s1m2a2 + … + s1mkak]
= l[s1a1 + s2a2 + ... + skak] = l(0) = 0.
M1 M 2 {0}
88 Modules
and this proves (*), that is, M is direct sum of M1 + M2. Now, M2 is generated by a2, a3, ..., ak. Thus, rank
of M2 is atmost k - 1. So by induction hypothesis, M2 is the direct sum of cyclic submodules. Therefore,
M is the direct sum of cyclic modules.
6.4.3. Schur's Lemma. If M is an irreducible or simple R-module, then the endomorphism ring EndR
(M) is a division ring.
Proof. Let ∅ (≠ 0) ∈ EndR(M), since ker∅ is a submodule of M and ker∅ ≠ M, so ker∅ = {0}, thus ∅ is
one one. Also, the image of ∅(M) is an R- submodule of M.
Since ∅ ≠ 0 => ∅(M) ≠ 0.
But M is irreducible => ∅(M) = M => ∅ is onto. Thus ∅ is one one and onto, hence has an inverse in
EndR (M). Hence, we have shown that every non zero element of EndR(M) is invertible. Hence EndR(M)
is a division ring.
6.4.4. Free Module. An R-module M is called a free module if M has a basis, that is, there exists a
subset S of M such that M is generated by S and S is linearly independent set.
6.4.5. Theorem. Let M be a free module with a basis {e1,e2,...en} then
M Rn
6.4.6. Annihilator of an element.
Let M be an R-module and x ϵ M. The subset {r ϵ R: rx = 0} is called annihilator of x and is denoted by
Ann(x).
6.4.7. Exercise. Ann(x) is a left ideal of R.
6.4.8. Theorem. Suppose 1 ϵ R, then M is a cyclic module iff M R I , where I is a left ideal of R.
Proof. First suppose that M is cyclic R-module, then M = Rx for some x ϵ M
Define a mapping f: R→M by f(r) = rx
Now we shall show that f is a homomorphism of left R- module
f(r + s) = (r + s)x = rx + sx = f(r) + f(s) r, s ϵ R
and f(sr) = (sr)x = s(rx) = sf(r) r, s ϵ R
Hence, f is an R-module homomorphism.
We shall prove that f is onto.
Let y ϵ M = Rx. Then, y = rx for some r ϵ R. Now, r ϵ R f(r) = rx = y. Hence, f is onto.
Thus, by Fundamental Theorem of Homomorphism R M , where I = ker f is a left ideal of R.
I
Conversely, suppose that R M
I
Let f: R M be the given homomorphism. Since 1+I ϵ R .
I I
Abstract Algebra 89
So let f(1+I) = x
Let y ϵ M. Now, f is onto, so there exists r+I ϵ R/I such that f(r+I) = y.
f(r (1+I)) = y r f(1+I) = y rx = y.
Hence y ϵ M y = rx ϵ Rx M Rx.
Again x ϵ M Rx M.
Hence, M = Rx
M is a cyclic module.
6.4.9. Theorem. Let R be a ring with unity and M be an R-Module. Then the following conditions are
equivalent:
(i) M is a simple R-module.
(ii) Every non-zero element of M generates M.
(iii) M R where I is a maximal ideal of R.
I
Proof. (i) (ii)
Let M be a simple R- module. Then, {0} and M are the only submodules.
Let x ≠ 0 ϵ M. Since 1 ϵ R and 1x = x ≠ 0, therefore x ϵ Rx = N = < x >
Now, N is a non zero submodule of M and M is simple, so N = M M = Rx, that is, every non zero
element of M generates M.
(ii) (i)
Suppose that M is generated by every non zero element of M. So let N ≠ {0} be a submodule of N. Take
0 ≠ x ϵ N. Then, by the given condition M = Rx
Now, x ϵ N Rx N, that is, M N. Also, N M M = N.
Thus the only non zero submodule of M is M itself. Therefore M must be simple.
(i) (iii)
Let 0 ≠ x ϵ M then N = Rx is a non zero submodule of M, since M is simple so N = M, that is, M = Rx.
Define a map f : R→ Rx = M by f(r) = rx, for all r ϵ R
It is easy to see that f is an onto homomorphism (as proved in previous theorem). Then by Fundamental
Theorem of Homomorphism for modules, we have
R M
ker f
Let I = ker f. Then, I is a module and R M .
I
90 Modules
We shall prove that I is maximal ideal of R.
Suppose J is a left ideal of R such that I J R . Now J is a submodule of R
I I
Since R M and M is simple, therefore, R is also simple. But then either J is zero submodule or
I I I
J = R , that is, J = {I} or J = R .
I I I I I
J = I or J = R.
Hence, I is a maximal left ideal of R.
(iii) (i)
Suppose M R where I is a maximal ideal of R.
I
Now, I is a maximal left ideal of R I ≠ R.
Now 1 ϵ R 1+I ϵ R and 1+I ≠ I (zero of R )
I I
By definition,
1(1+I) = 1 1+I = 1+I ≠ I R(1+I) ≠ I R( R ) ≠ I RM ≠ {0}
I
Let N be a submodule of M. Since M R , so there exists some isomorphism : M R . Then
I I
(N) is a submodule of R and precisely of the form J where J is a left ideal of R containing I.
I I
Now I is maximal left ideal of R and I J R.
either J = I, then J = I = zero element of R .
I I
(N) = {I}
But is one one N = {0}
If J = R, then (N) = R M N = M
I
Hence {0} and M are the only submodules of M.
Therefore, M must be simple.
6.4.10. Definition. Let M be an R-module and N be a submodule. We say that N is a direct summand of
M if there exists another submodule N ' of M such that
M=N N' .
Abstract Algebra 91
6.4.11. Theorem. Let M be an R-module and N be a submodule of M such that M is a free R-module
N
then N is a direct summand of M.
Proof. Let {x N} be a basis of M over R, where x M . Since a basis is linearly
N
independent, therefore this basis will not contain zero element.
Thus, x N N , that is, x N .
Let B = {x } M and M ' x
, that is, M ' is a submodule of M generated by B. We shall
prove that M = N M '
Let x ϵ M, then x + N ϵ M . Now {x N} is a basis of M , so there exists r1, r2, …, rn ϵ R such
N N
that
x+N = r1 (x 1 N) r2 (x 2 N) . . . rn (x n N)
= (r1 x 1 N) (r2 x 2 N) . . . (rn x n N)
= (r1 x 1 r2 x 2 . . . rn x n ) N
n
= r x N
i 1
i i
n
x ri xi N
i 1
n n
Let x ri xi y, y N x y ri xi N M .
i 1 i 1
Thus M = N + M' . Let Z ϵ N ∩ M' , then Z ϵ N & Z ϵ M' . Now Z ϵ M' =<B>
Z t1 x1 t2 x2 ... tn xn , ti R
n n
Z + N = ti xi N ti ( xi N)
i 1 i 1
n
ti ( xi N) Z + N = N [ Z N] .
i 1
Since {x N} is a basis of M , therefore, ti = 0 i 1 i n
N
n
But then Z = ti xi 0 N M' = {0}
i 1
Now it is clear that M=N M' , that is, N is direct summand of M.
92 Modules
6.4.12. Theorem. Let M1, M2 be free submodules of M such that M1 + M2 = M1 M2 , then M1+M2 is
also a free R-module.
Proof. Now M1 is a free R-submodule of M, so M has a basis say {xi }iI M1 .
Similarly, M2 has a basis say {y j } jJ M 2
Let B = {x i }iI {y j } jJ
Then, B M1+M2 [ M1 ,M2 M1 +M2 ]
Let x ϵ M1+M2, then x = x' + y' where x' M1 and y' M2 . Now {xi }iI is a basis of M1 over R.
x' is a linear combination of finite number of elements, say x1, x2, . . ., xm, that is,
m
x' = rk xk , rk R .
k 1
Similarly, there exist elements y1, y2, . . ., yn such that
n
y'= st yt , st R
t 1
m n
Therefore, x x ' y ' rk xk st yt .
k 1 t 1
M1 + M2 is generated by B.
Next, suppose B = {zk }kK . Let r1, r2, . . ., rn be such that r1z k1 + r2 z k 2 + . . . + rn z kn =0 , zi R .
Since ‘+’ is commutative, we may assume that z k1 , z k 2 , . . . , z km {x i }iI and
z k m+1 , z k m+2 , . . . , z k n {y j } jJ .
Hence, r1z k1 + r2 z k 2 + . . . + rm z k m = -(rm+1z k m+1 + rm+2 z km+2 + . . . + rn z k n ) .
r1z k1 + r2 z k2 + . . . + rm z k m ∈ M1 and -(rm+1z k m+1 + rm+2 z k m+2 + . . . + rn z kn ) ∈ M2.
Hence, r1z k1 + r2 z k 2 + . . . + rm z k m M1 M 2 = {0}, that is, r1z k1 + r2 z k2 + . . . + rm z km =0 .
Similarly rm+1z k m+1 + rm+2 z k m+2 + . . . + rn z k n = 0.
Since {xi }iI and {y j } jJ are linearly independent. It follows that ri = 0, 1 ≤ i ≤ n
Hence B is linearly independent.
From this it follows that B is a basis for M1 + M2 = M1 M2 over R.
Hence M1 + M2 is a free R-modules.
Abstract Algebra 93
6.4.13. Theorem. Let R be a commutative ring with unity and e be an idempotent (that is, e2 = e and
e ≠ 0, e ≠ 1). Let M = Re, then M is not a free R-module.
Proof. Since e ≠ 0, e ≠ 1. Thus {0} M R.
Suppose M is a free R-module. Since M = R e, so M is finitely generated R-module. Thus every basis of
M has finite number of elements. Let {x1, x2, . . ., xm} be a basis of M. now xi∈ M = R e.
x i = ri e , 1 i m, ri R .
Since {x1, x2, . . ., xm} is a basis of M, so each xi ≠ 0 ri 0
Now, r2x1+(-r1)x2 = r2r1e + (-r1)r2e
= r2r1e - r1r2e = 0 [ R is commutative]
Now - r1 0, r 2 0 x1 and x2 are linearly dependent, a contradiction. Therefore, the basis has only
one element, namely x1 = r1e.
Now, e ≠ 0, 1 1 – e ≠ 0. 1-e ∈ R, so (1 - e)x1 = (1 - e)r1e = r1(e – e2) = r1(e - e) = 0.
But 1 – e ≠ 0 and so x1 is linearly dependent, again a contradiction.
Thus, M is not a free R-module.
6.4.14. Theorem. Let N be a finitely generated free module over a commutative ring R. Then all its
basis are finite.
Proof. Suppose N is generated by {x1, x2,…, xn} and {ei }i be a basis of N and let us denote this basis
by B. We shall prove that B is a finite set.
Now, xi N and B is a basis of N, so there exists a finite subset Bi of B such that xi is a linear
combination of elements of Bi with coefficients in R.
n
Let S = B , then clearly S is finite since each Bi is finite.
i 1
i
Now, B is linearly independent, so S, being a subset of B, is also linearly independent. Let x N be
any arbitrary element, then
x = r1 x1+ r2 x2 +… + rn xn, ri R
But each xi is a linear combination of elements of Bi, so x is a linear combination of elements of S.
Hence S generates N. Thus, S is a basis of N. Now S B and B is also a basis, so we must have
S = B. herefore, B is finite since S is finite.
6.4.15. Theorem. Let N be a finitely generated free module over a commutative ring R. Then all basis
of N have the same number of elements.
Proof. Suppose N has two basis containing m and n elements respectively. We shall prove that m = n.
Since N is a free module, so we must have (by a previous theorem), that N R m and N R n
R m Rn .
94 Modules
Let, if possible, m < n.
Let : R m R n be an isomorphism and since is one-one and onto, so is invertible and let =
–1, then : R n R m is an isomorphism.
Let {e1, e2,…, en} and { f1, f2,…, fn} be standard basis of R m and R n respectively.
Now ei R n , 1 i m, so let us write
(ei) = a1i f1 + a2i f2 + … + ani fn, 1 i m
that is, we have
(e1) = a11 f1 + a21 f2 + … + an1 fn
(e2) = a12 f1 + a22 f2 + … + an2 fn
. . .
(em) = a1m f1 + a2m f2 + … + anm fn
Then, matrix of = A =
Again, (fi) R m , 1 j n, so let us write
(fi) = b1j e1 + b2j e2 + … + bmj em
that is, we have
(f1) = b11e1 + b21e2 + … + bm1em.
(f2) = b12e1 + b22e2 + … + bm2em
: : :
(fn) = b1ne1 + b2ne2 + … + bmnem
b11 b12 b1m
b b22 b2 m
Matrix of = B =
21
bm1 bm 2 bmn mn.
Now, we see that : R m R n , : R n R m implies that : R m R m and is identity
mapping on R m because = –1.
So matrix associated with the mapping is identity matrix, but matrix of is also given by BA.
Hence, BA = Im = Identity matrix of m m.
Abstract Algebra 95
that is,
m
1 if i j
b ik akj =
i j
(1)
k 1 0 if
Similarly, AB = In, implies
m
1 if i j
a ik bkj =
i j
(2)
k 1 0 if
B
Let A = [A O] and B= be n n augmented matrices, where each of the O blocks is a matrix of
O
appropriate size, that is,
Form these two matrices we note that
So, we have
I 0
det(AB) = det (In) = 1 and det (BA) = det m = 0.
0 0
det (AB) det (BA) …(3)
96 Modules
But A and B are n n matrices over a commutative ring so we must have det(AB) = det(BA), which
is contradicted by (3).
Hence m n.
Similarly n m.
So, m = n, that is, all basis of N have the same number of elements.
6.4.16. Rank. The number of elements in any basis of a finitely generated free module N over a
commutative ring R with unity is called the rank of N.
6.4.17. Theorem. Every finitely generated module is a homomorphic image of a finitely generated free
module.
Proof. Let N be an finitely generated R-module with generators x1, x2, …, xn.
Let e1 = (0, 0, …, 1, 0, . . ., 0) be the n-tuple with all entries 0 except the ith place, where the entry is
1. Then, we know that {e1, e2, …, en} are linearly independent over R and generated a free module
R n . Hence R n is a finitely generated free module. We shall prove that N is homomorphic image of
R n . We define a mapping
n n
: R n N by setting rei i rx i i
i 1 i 1
n n
(i) is well-defined. Let x = rei i
i 1
and y = r 'e
i 1
i i be two elements of R n such that x = y,that
n
is (r r ' ) e
i 1
i i i 0 ri ri' 0 , 1 i n [Since ei' are L. I.]
n n
ri ri' , 1 i n ri xi
i 1
r 'x
i 1
i i ( x) ( y ) .
(ii) is homomorphism.
n n
Let x = re
i 1
i i ,y= r 'e
i 1
i i and r R, then
n n
(x + y) = (ri ri' ) ei (r r ' ) x
i i i
i 1 i 1
n n
= r x r 'x
i 1
i i
i 1
i i = (x) + (y)
n n
and (rz) = rrei i = rr x i i r( x) .
i 1 i 1
Abstract Algebra 97
n
(iii) is onto. Let m = rx
i 1
i i be any arbitrary element. Then ri R and consider the element
n
n n
x= re i i ri xi = m. So, x is pre-image of m. Hence is onto
n
i i R . Then (x) = re =
i 1 i 1 i 1
and so, ( R n ) = N, that is, N is homomorphic image of R n which is a finitely generated free-module.
6.4.18. Theorem. Every finitely generated module is isomorphic to a quotient group of a finitely
generated free module.
Proof. Proof can be obtained by using Theorem 6.4.14 and then using fundamental theorem of module
homomorphism
R n ker M.
Thus, N is isomorphic to a quotient group of finitely generated free-module.
6.4.19. Fundamental Structure Theorem (or Decomposition Theorem) of finitely generated
modules over Principal Ideal Domain.
Let R be a PID and let N be any finitely generated R-module, then
N R s R Ra1 R Ra2 ... R Rar
a direct sum of cyclic modules, where ai' s are non zero non-units and ai ai 1 , i = 1, 2,.., r 1.
Proof. Since N is a finitely generated R-module and we know that every finitely generated module is
isomorphic to a quotient group of a finitely generated free module, so
M Rn K .
Now, since R n is a free R-module, where R is a PID and K is a sub module of R n , so we must have
K R m , where m n.
Let be this isomorphism from R m to K, that is, K = ( R m ). Let { e1, e2, …, em} be a basis of R m .
Let us write
a11
a
(e1 ) 21 R n
an1
a1m
a
(em ) 2 m R n
anm
98 Modules
Then, ( R m ) = A R m , where A = (aij) is an n m matrix. We choose invertible matrices P and Q of
order n n and m m respectively such that
PAQ = diag.((a1, a2, …, ak, 0, 0, …, 0), where a1 a2 ... ak .
Then we have
M Rn K R n (R m ) R n AR m R n PAQR m
a1
a2 O
R
R
Rn ak
0
O R
0
R R ... R Ra1 Ra2 ... Rak
R Ra1 R Ra2 ... R Rak R
...
R
(n --k ) times
By deleting the zero terms on R.H.S., if any (corresponding to those ai's that are units) and renumbering
if necessary we obtain
M R Ra1 R Ra2 ... R Rar R s .
6.4.20. Application to finitely generated abelian groups. Since the ring of integers is a PID and
any abelian group is a module, so by above theorem it follows that:
Let A be a finitely generated abelian group, then
A s a1 ... ar
where ‘s’ is a non-negative integer and ai's are non-zero non units in s.t. a1 a2 ... ar .
6.5. Torsion Element. Let M be an R-module, then an element m of M is said to be a torsion element if
rm=0 for some nonzero element r ∈ R. The collection of all torsion elements is denoted by Mt or
Tor(M), so
Tor(M) = {m ∈ M : rm=0 for some non-zero r ∈ R}.
Also, it is known as Torsion part of the module. It is the largest torsion submodule of M.
65.1. Exercise. Prove that if R is an integral domain, then Tor(M) is a submodule of M.
6.5.2. Torsion module. A module is said to be a torsion module if its every element is a torsion element.
Abstract Algebra 99
6.5.3. Torsion free moldule. A module having no non-zero torsion elements is called a torsion-free
module or we can say that the only torsion element in a torsion-free module is the zero element, which
means that every non-zero element of this module is linearly independent.
Now according to the definition of Torsion part of a module, we can say that M is torsion free if its
torsion part Mt is {0}.
6.5.4. Proposition. For any module M over a commutative integral domain R, the quotient module
M/Mt is a torsion free module.
Proof. Let x + Mt ∈ M/Mt be any arbitrary element. If x + Mt is a torsion element of M/Mt, then r.(x +
Mt) = 0 for some non-zero r ∈ R and, therefore, r.x ∈ Mt. So by definition of torsion part of M, there
exists some non-zero element a ∈ R such that a(r.x) = 0. Then, (ra).x = 0, since R is an integral domain
and a, r are non-zero, so ra is also non-zero element of R. Therefore, x is a torsion element of M and thus
x ∈ Mt. Hence x + Mt = Mt , which is zero element of M/Mt. Hence zero element is the only torsion
element of M/Mt and so M/Mt is a torsion free module.
6.5. Check Your Progress.
1. Let M1, M2 be free R-module. Then M=M1×M2 is also a free module.
2. Q is not free Z-module.
Answers.
1. First prove M1×M2 M1 M2 . Since direct sum of two free modules is again free module, the result
follows.
2. Let M = Q. We know that every abelian group can be treated as Z-module. Thus, Q is a Z-module.
Suppose M is a free Z-module. Let {x λ } be a basis of M over Z. Let x1, x2∈ {x λ } . Suppose
m m
x1 = 1 , x 2 = 2 where m1, m2, n1, n2∈ Z and n1 ≠ 0, n2 ≠ 0.
n1 n2
m1 m
Now (n1m2)x1 + (-n2m1)(x2) = n1m2 n 2 m1 2 = m1m2 m1m2 = 0, n1m2 0, n 2 m1 0. Thus x1,
n1 n2
x2 are linearly dependent, a contradiction. Hence, basis of M contains only a single element, say x.
m1
Suppose x = , where m1, n1∈ Z. and n1 ≠ 0.
n1
Now {x} is a basis of M over Z. Thus, every element of M is of the form lx, l ∈ Z.
Take a prime p such that p > n1
1
Now, ∈M=Q
p
1 m
l 1 for some l ∈ Z.
p n1
100 Modules
n1 lm1 p p n1 , a contradiction. ( p > n1)
Hence, M = Q is not a free Z-module.
6.6. Summary. In this Chapter, we discussed about various properties of modules, structures that
become module, their direct sum, rank etc. Also, it was derived that for a finitely generated free module
over a commutative ring all basis have same number of elements.
Books Suggested:
1. Luther, I.S., Passi, I.B.S., Algebra, Vol. I: Groups, Vol. III: Modules, Narosa Publishing House
(Vol. I – 2013, Vol. III –2013).
2. Lanski, C. Concepts in Abstract Algebra, American Mathematical Society, First Indian Edition,
2010.
3. Sahai, V., Bist, V., Algebra, Narosa Publishing House, 1999.
4. Malik, D.S., Mordenson, J.N. and Sen, M.K., Fundamentals of Abstract Algebra, McGraw Hill,
International Edition, 1997.
5. Bhattacharya, P.B., Jain, S.K. and Nagpaul, S.R., Basic Abstract Algebra (2nd Edition),
Cambridge University Press, Indian Edition, 1997.
6. Musili, C., Introduction to Rings and Modules, Narosa Publication House, 1994.
7. Jacobson, N., Basic Algebra, Vol. I & II, W.H Freeman, 1980 (also published by Hindustan
Publishing Company).
8. Artin, M., Algebra, Prentice-Hall of India, 1991.
9. Macdonald, I. D., The Theory of Groups, Clarendon Press, 1968.
7
NOETHERIAN AND ARTINIAN MODULES
Structure
7.1. Introduction.
7.2. Noetherian and Artinian Module.
7.3. Hilbert Basis Theorem.
7.4. Nil Ideal.
7.5. Ring of Homomorphisms.
7.6. Radical ideal.
7.7. Check Your Progress.
7.8. Summary.
7.1. Introduction. In this chapter the concept of ascending and descending sequences of submodules,
definitions, examples and properties of Noetherian and Artinian Modules are given. Further nil ideal and
nilpotent ideals are discussed in detail.
7.1.1. Objective. The objective of these contents is to provide some important results to the reader like:
(i) Noetherian Module and equivalent conditions for a module to be Noetherian.
(ii) Artinian Module and equivalent conditions for a module to be Artinian.
(iii) Hilbert Basis Theorem.
(iv) Wedderburn Artin Theorem.
7.1.2. Keywords. Noetherian Module, Artinian Module, Nil Idels, Finitely Generated Module.
7.2. Noetherian and Artinian Modules. Let M be a left R-module and {Mi }i ≥ 1 be a family of
submodules of M
1) The family {Mi }i ≥ 1 is called an ascending chain or sequence if
M1 M 2 M 3 ... M n M n1 ...
2) The family {Mi }i ≥ 1 is called a descending chain or sequence if
M1 M 2 M 3 ... M n M n1 ...
102 Noetherian and Artinian Modules
7.2.1. Noetherian Module.
An R-Module M is called Noetherian if for every ascending chain
M1 M 2 M 3 ... M n M n1 ...
of submodules of M, there exists a positive integer K such that
MK = MK+1 = MK+2 = . . .
that is, MK = MK+i, for all i ≥ 0.
-OR-
Every ascending chain of submodules of M becomes stationary after a finite number of steps.
-OR-
Every properly ascending chain M 1 M 2 M 3 ... of submodules of M terminates after a finite number
of steps, that is, every properly ascending chain of sub modules of M is finite.
7.2.2. Theorem. For an R-module M the following conditions are equivalent:
(i) M is Noetherian.
(ii) Every non-empty family of R-modules has a maximal element.
(iii) Every sub-module of M is finitely generated.
Proof : (i) (ii)
Suppose M is Noetherian R-module and let = M be a non-empty family of sub-modules of M.
Since the family is non-empty so let M 1 be any member of .
If M 1 is maximal element, then we are done, otherwise there exist M 2 such that
M 1 M 2 .
Again, if M 2 is maximal, then we are through, otherwise there exist M 3 such that
M 2 M3 .
Now, has no maximal element is equivalent to saying that there exists an infinite ascending chain
M 1 M 2 M 3 ...
of sub-modules of M, which is a contradiction as M is assumed to be Noetherian. Hence the family
must have a maximal element.
(ii) (iii)
Abstract Algebra 103
We are given that every non-empty family of sub-modules of M has a maximal element. Let N be a sub-
module of M and we shall prove that N is finitely generated.
Let, if possible, N is not finitely generated. For any positive integer k let a1, a2,…, ak N.
Then N < a1, a2,…, ak >. Choose ak+1 N such that ak+1 < a1, a2,…, ak >.
We then obtain an infinite properly ascending chain
< a1 > < a1, a2 > ... < a1, a2,…, ak > ...
of submodules of M.
Let us denote Nk = < a1, a2,…, ak >, then N1 N2 … Nk ...
Let family of all these sub-modules be , that is, = {Nk}, k 1, then by the given hypothesis has
a maximal element, say L.
Now, L L = Nm for some m. But L = Nm Nm +1 so L is not a maximal element, a
contradiction. Hence N must be finitely generated.
(iii) (i)
Suppose every submodule of M is finitely generated. We shall prove that M is Noetherian.
Let N1 N2 N3 ... Nk Nk+1 ... be an ascending chain of sub-modules of M.
Consider N = N
i
i . We claim that N is also a sub-module of M.
Let x, y N
i
i and r R. Then x Nr and y Ns for some integers r and s.
Since either Nr Ns or Ns Nr, therefore both x and y lie in one sub-module Nr or Ns, and hence x y
and rx lie in same sub-module. However, both Nr and Ns are subsets of N, so x y, rx N, and hence N
is a submodule of M.
Now, by (iii), N is finitely generated. So there exist elements a1, a2,…, an N such that
N = < a1, a2,…, an >. Now for each j, 1 j n, aj N = N
i
i
aj N j for some natural number j
Let k = max 1 , 2 ,..., n , then clearly aj Nk, 1 j n that is,, a1, a2,…, an Nk
But N = < a1, a2,…, an >, so N is smallest submodule containing a1, a2,…, an .
This implies N Nk. . Also Nk N. Hence N = Nk and so Nk = Nk +1 = Nk +2 = ..., therefore, M is a
Noetherian R-module.
104 Noetherian and Artinian Modules
7.2.3. Definition. If the left R-module M is noetherian then M is called a left-Noetherian module.
Similarly, if a right R-module M is Noetherian then M is called a right Noetherian module.
7.2.4. Finitely Co- generated Module.
Let M be a left R- module then M is called finitely co- generated if for any non- empty family {M } £
of sub modules of M, having {0} intersection, that is, if M {0} . Then, there exists λ1, λ2, …, λn
n
such that M
i 1
i
{0}.
7.2.5. Artinian Module.
Let M be a left R- module. M is called left- Artinian if for every decreasing chain
… …
we have ⋯ ⋯ for some k N, that is, have for all i 0, that is,
every descending chain of sub modules of M becomes stationary after a finite number of steps.
OR
Every properly descending chain of M 1 M 2 M 3 ... M n ... of submodules of M is finite.
OR
Every properly descending chain of sub modules of M terminates after a finite number of steps.
7.2.6. Theorem. For an R-module M the following conditions are equivalent:
(i) M is Artinian.
(ii) Every non-empty family of sub-modules of M has a minimal element.
(iii) Every quotient module of M is finitely co-generated.
Proof. (i) (ii)
Let us suppose that M is Artinian and let = M be a non-empty family of sub – modules of M.
We shall prove that has a minimal element. Now so there exist M 1 , then either M 1 is
a minimal element of or there exists M 2 such that M 1 M 2 .
Again, either M 2 is a minimal element of or there exists M 3 such that
M 2 M3 .
If this process continues indefinitely (that is, has no minimal element) then we get an infinite
properly descending chain M 1 M 2 M 3 ... of sub-modules of M, which is a contradiction since M
is given to be Artinian. Hence must have a minimal element.
Abstract Algebra 105
(ii) (iii)
Let us suppose that every non-empty family of sub-modules of M has a minimal element. We shall
prove that every quotient module of M is finitely co-generated. Let N be a sub-module of M. Consider
the quotient module M N .
Let M N be a family of sub-modules of M N such that M N = {N}
Now, {N} = M N = M N M =N (1)
Let = M , where M ’s are sub-modules of M and let
= A : A is the intersection of finite number of sub-modules of M in
Then clearly that is,, M .
Now is a family of sub-module of M so by the given condition (ii) it must have a minimal element
say, A.
Then, A = M 1 M 2 ... M n , i
Let M be any member, then A M = M M 1 M 2 ... M n , being finite
intersection of members of .
Now A M A. But A is minimal element of so A M = A
A M
A M =N [By (1)]
n n
Again, N = M M i = A . Hence A = N = M i
i =1 i =1
n
M
n
Now, i N = M i N = N N = {N}.
i =1 i =1
n
Hence, there exist a finite sub-family M i N
i =1
of M N such that
M
i =1
i
N = {N}
Hence every quotient module of M is finitely co-generated.
(iii) (i)
106 Noetherian and Artinian Modules
Let us suppose that every quotient module of M is finitely co-generated and shall prove that M is
Artinian. Let M1 M2 ... Mk Mk +1 ... be a descending chain of sub-modules of M.
Let N = M i . Then N is a sub-module of M and N Mi for all i. Consider the family M i Ni of
i
sub-modules of M N . We see that
M
i
i N = Mi N = N N = {N}
i
Since by the given condition (iii), M N is finitely co-generated therefore there exists a finite sub-
r r
family, say, M n i N of M i Ni such that M ni N = {N}
i =1
i =1
Let k = max. {n1, n2,...,nr} then
r
M
i 1
ni Mk ,
since the chain is descending.
r
r
Now {N} = M
i 1
ni
N Mn i N M k N and so Mk = N.
i 1
But then N M k j M k N gives that Mk = Mk +1 = Mk +2 = ...
Hence, the R-module M is Artinian.
7.2.7. Proposition. Let M be a left R-Module
1) If M is Noetherian then every submodule and factor module of M is also Noetherian.
2) If N is a submodule of M such that bot h N and are Noetherian then so is M.
Proof. 1) We know that “a R-module M is Noetherian iff every sub-module of M is finitely generated.”
(i) Let N be a sub-module of M then N must be finitely generated. Now let N1 be any sub-module of N
then it is also finitely generated as every sub-module of a finitely generated module is finitely generated.
Hence N is Noetherian.
(ii) Let M N be any quotient module of M. To prove M N Noetherian, we shall prove that every sub-
module of M N is finitely generated. So let A N be any submodule of M N , where A is a sub-
module of M. Now A is a sub-module of M and M is Noetherian, therefore A is finitely generated.
Suppose A = < x1, x2,...,xn > .We claim that A N = < x1 + N, x2 + N, ..., xn + N >
Let x + N A N be any element.
Then, x A x = r1x1 + r2 x2 +...+ rn xn, ri R
x + N = (r1x1 + r2 x2 +...+ rn xn) + N
Abstract Algebra 107
= r1 (x1+N) + r2 (x2 +N) +...+ rn (xn + N)
Hence A N = < x1+N, x2 +N, ...,xn+N >, that is, A N is finitely generated. Therefore, M N is
Noetherian.
2) To prove that M is Noetherian, we shall prove that every sub-module of M is finitely generated. So,
let A be any sub-module of M. Then A + N is also a submodule of M containing N. Therefore A+N
N
is a sub-module of M N . Since M N is Noetherian, therefore, A+N must be finitely generated.
N
Now, we know that A+N A AN
N
A A N is also finitely generated.
Also, A N being a sub-module of Noetherian R- module N, is also finitely generated.
Suppose A N = < x1, x2,..., xm > and A A N = < y1 + A N, y2 + A N,..., yn + A N >,
where xi, yi A.
We claim that A = < x1, x2,...,xm, y1, y2,...,yn >
Let x A be any arbitrary element, then
x + A N A A N and A A N = < y1+A N, y2 + A N,...,yn+A N >
n
so x + AN = r y
j 1
j j A N , rj R
n
= r y
j 1
j j AN
n
x ry j j AN AN
j 1
n
x rj y j A N .
j 1
n
Since A N = < x1, x2,...,xm >, so x rj y j = s1x1 + s2x2 +...+ smxm, si R
j 1
n m
x rj y j
j 1
sx
i 1
i i
A = < x1, x2,...,xm, y1, y2,...,yn >
Hence A is finitely generated and therefore M is Noetherian R-module.
7.2.8. Proposition. Let M be a left R-module and N be a sub module of M. Then M is Artinian iff both
N and are Artinian.
Proof. First suppose that both N and are Artinian.
108 Noetherian and Artinian Modules
We shall prove that M is Artinian.
Let A1 A2 ... An An +1 be a descending chain of sub-modules of M. We shall prove that this
chain becomes stationary after a finite number of steps.
Now, Ai is a sub-module of M and N is also a sub-module, so Ai + N is also a sub-module of M and N
Ai + N for all i.
Since, Ai Ai +1 so Ai + N Ai+1+ N for all i.
Ai N N Ai +1 N N ,
where Ai N N is a sub-module of M N for all i.
Hence, we have a descending chain A1 N N A2 N N ... An N N ... of submodules of
M N.
Since M N is Artinian, so there exist a positive integer r such that
Ar N N = Ar +1 N N = Ar +2 N N = ...
that is, Ar N N = Ar +i N N i 0
Ar + N = Ar + i N i 0 (1)
Now, again Ai N is a sub-module of N and since Ai Ai+1.
We have Ai N Ai 1 N for all i. Hence, we have a descending chain
A1 N A2 N ... An N ...
of sub-modules of N. However, N is artinian, so there exist a positive integer s such that
As N = As +1 N = As +2 N = ...
that is,, As N = As+i N for all i 0. (2)
Let k = max{r, s}, then by (1) and (2), we must have
Ak +N = Ak + i + N for all i 0 (3)
and Ak N = Ak+i N for all i 0. (4)
We claim that Ak = Ak + i for all i 0.
Let x Ak be any arbitrary element. Then x Ak Ak +N = Ak+i + N [By(3)]
x = y + z, for some y Ak+i, z N
x y = z N.
Now, x Ak, y Ak+i Ak x y Ak
Hence x y Ak N = Ak+i N [By (4)]
Abstract Algebra 109
x y Ak+i N Ak+i
x y Ak+i and y Ak+i
x y + y = x Ak+i.
So, Ak Ak+i, also Ak+i Ai.
Hence, we have Ak = Ak+i for all i 0.
Therefore, the considered chain of sub-modules of M becomes stationary after a finite number of steps.
Hence M is Artinian.
Conversely, suppose that M is an Artinian module.
(i) Let N be any sub-module of M. We shall prove that N is Artinian.
Suppose that N1 N2 N3 ... Nk Nk +1 ... be a descending chain of sub-modules of N. Since
sub-modules of N are also the sub-modules of M, it follows that above chain is a descending chain of
sub-modules of M. Since M is Artinian, therefore, there exists a positive integer k such that
Nk = Nk +1 = Nk +2 = …
Hence N is Artinian.
(ii) Let M N be any quotient module of M. To prove M N Artinian, let us consider a descending chain
of sub-modules of M N , that is,
M1 N M2 N M3 N ... Mk N ...
Here, Mi are sub-modules of M and since Mi N Mi +1 N , so Mi Mi +1.
Thus, we have a descending chain M1 M2 ... Mk ... of sub-modules of M. Since M is Artinian,
so there exists a positive integer k such that Mk = Mk +1 = Mk +2 = ... and then, we have
Mk N = Mk +1 N = Mk +2 N = ...
Hence, M N is a Artinian R-module.
7.2.9. Remark. Since every homomorphic image of a module is isomorphic to some quotient module.
Thus, if the module is Noetherian (Artinian), then its homomorphic image is also Noetherian (Artinian).
7.2.10. Result. Let M1 and M2 be R-modules and N1 and N2 be submodules of M1 and M2 respectively.
Let M = M1 M2 = {(x, y) : x M1 , y M 2 }. Define addition and scalar product as
(x1,y1) + (x2,y2) = (x1 + x2, y1 + y2)
r(x, y) = (rx, ry).
Then, M is a module with respect to this addition and scalar multiplication.
Also, let N = N1 N2. Then, N is a submodule of M. Consider the factor module
110 Noetherian and Artinian Modules
M M1 M 2
.
N N1 N 2
M1 M2 M1 M2
Consider the quotient module and and let be their direct product. Define a
N1 N2 N1 N2
M1 M2
map : M 1 M 2 by
N1 N2
( x, y) ( x N1 , y N2 ) .
It is easy to see that is the homomorphism of left R-modules.
M1 M2
Now, is onto. For, let ( x N1 , y N 2 ) , where x M1 and y M 2 . Therefore,
N1 N2
( x, y) M1 M 2 and ( x, y) ( x N1 , y N2 ) .
Also, let ( x, y) Ker M1 M 2 . Then,
( x, y ) zero of M 1 N M 2 N ( N1 , N 2 ) iff ( x N1 , y N2 ) ( N1 , N2 )
1 2
iff x N1 N1 and y N2 N2 iff x N1 and y N 2
Ker N1 N2 .
M1 M 2 M1 M2
Hence, by fundamental theorem of homomorphism .
N1 N 2 N1 N2
Generalising this we get
M 1 M 2 ... M n M1 M2 Mn
... .
N1 N 2 ... N n N1 N2 Nn
M1
Remark. If M1 is a left R-module, then {0}.
M1
7.2.11. Proposition. Let M1, M2, ..., Mk be left R-modules.
(i) If each of Mi is Noetherian, then so is M1 M 2 ... M k .
(ii) If each of Mi is Artinian, then so is M1 M 2 ... M k .
Proof. We shall prove the result by induction on k.
Suppose k = 2. We know that "If N is a submodule of M, then M is Noetherian iff both N and M/N are
Noetherian."
Now, M 1 M 2
M1
M2
{0}
M2
M2
M 2.
M 1 {0} M1 0 0 0
Abstract Algebra 111
M M2
Since M 2 is Noetherian, therefore 1 is Noetherian.
M 1 {0}
M M2
Now both M1 {0} and 1 are Noetherian, therefore, M1 M 2 is also Noetherian.
M 1 {0}
Suppose k > 2 and the result holds for modules M1, M2, ..., Mk-1.
Now M1 M 2 ... M k 1 is Noetherian and Mk is also Noetherian, therefore by above discussion
M1 M 2 ... M k is Noetherian. Hence the Proof.
k
7.2.12. Proposition. Let M1, M2, ..., Mk be Noetherian submodules of M, then M
i 1
i is also
Noetherian.
Proof. We know that finite direct product of Noetherian Modules is again Noetherian. Now, each of M1,
M2,..., Mk is Noetherian, therefore, M1 M 2 ... M k is Noetherian.
k
Define a map : M1 M 2 ... M k M i by
i 1
( x1 , x2 ,..., xk ) x1 x2 ... xk , xi M i .
Now, is a homomorphism:
Let ( x1 , x2 ,..., xk ) , (y1 , y2 ,..., yk ) M1 M 2 ... M k . Then,
(( x1 , x2 ,..., xk ) (y1 , y2 ,..., yk )) ( x1 y1 , x2 y2 ,..., xk yk )
k k k
( xi yi ) xi yi
i 1 i 1 i 1
( x1 , x2 ,..., xk ) (y1 , y2 ,..., yk ).
k k
and (r( x1 , x2 ,..., xk )) (r x1 , r x2 ,..., r xk ) rxi r xi r ( x1 , x2 ,..., xk ).
i 1 i 1
Hence, is a homomorphism.
is onto:
Let x M1 M 2 ... M k . Then, x = x1 + x2 +...+ xk, xi M i .
Now, ( x1 , x2 ,..., xk ) M1 M 2 ... M k and ( x1 , x2 ,..., xk ) x1 x2 ... xk x.
Thus, is onto.
k
Hence, M
i 1
i is a homomorphic image of a Noetherian module M1 M 2 ... M k .
112 Noetherian and Artinian Modules
Therefore, M1 M 2 ... M k is Noetherian.
Exercise. If the submodules M1 , M 2 ,..., M k are Artinian, then so is their sum.
7.2.13. Left Noetherian Ring. Let R be a ring. R is called left Noetherian, if left R-module RR is
Noetherian.
7.2.14. Right Noetherian Ring. Let R be a ring. R is called right Noetherian, if right R-module RR is
Noetherian.
7.2.15. Noetherian Ring. If R is both left as well as right Noetherian, R is called Noetherian ring.
7.2.16. Remark. Suppose R is a commutative ring, then it is clear that R is left Noetherian
iff R is right Noetherian iff R is Noetherian.
7.2.17. Artinian Ring. Let R be a ring. Then,
(1) R is called left Artinian if Left R module RR is Artinian.
(2) R is called right Artinian if right R module RR is Artinian.
(3) If R is both left as well as right Artinian, R is called Artinian ring.
(4) If R is a commutative ring, then it is clear that R is left Artinian iff R is right Artinian iff R is
Artinian.
7.2.18. Proposition. Let R be a right Noetherian ring and M be a finitely generated R-module, then M is
also Noetherian.
Solution. We know that finite direct product of Noetherian left R-Modules is again Noetherian,
therefore finite direct product of RR 's is also Noetherian.
Therefore, M R R R R ... R R R( n ) is Noetherian.
Now, M is finitely generated R-module, thus there exist x1 , x2 ,..., xn M such that
n
M Rxi x1 , x2 ,..., xn
i 1
Now, R( n ) M {(r1 , r2 ,..., rn ) : ri R} is Noetherian R-module.
Define a map : M M by
(r1 , r2 ,..., rn ) r1 x1 r2 x2 ... rn xn .
Then, is a homomorphism:
Let (r1 , r2 ,..., rn ) M and (s1 ,s2 ,...,sn ) M . Then,
Abstract Algebra 113
((r1 , r2 ,..., rn ) (s1 ,s2 ,...,s n )) (r1 s1 , r2 s 2 ,..., rn s n )
n n
ri x i si x i
i 1 i 1
(r1 , r2 ,..., rn ) (s1 ,s 2 ,...,s n ).
Let r R R , then
n n
(r (r1 , r2 ,..., rn )) (rr1 , rr2 ,..., rrn ) rri xi r ri x i r (r1 , r2 ,..., rn ).
i 1 i 1
Hence, is a homomorphism of left R-modules.
n
Let x M , then, there exists r1 , r2 ,..., rn R such that x ri x i .
i 1
n
But then (r1 , r2 ,..., rn ) R( n ) and (r1 , r2 ,..., rn ) ri x i x.
i 1
Hence, is onto.
Thus, : M M is epimorphism that is, M is homomorphic image of M' and so, M is Noetherian.
7.2.19. Exercise. Let R be a left Artinian ring and M be a finitely generated left R-module, then M is
also Artinian.
Proof. Proof of this exercise is similar to that of the above if we change Noetherian by Artinian.
7.3. Hilbert Basis Theorem. Let R be a left Noetherian ring then so is the polynomial ring R[x] and
conversely.
Proof. Let the ring R be left Noetherian and let R[x] be the polynomial ring over R in the indeterminate
x.
We shall prove that every left ideal of R[x] is finitely generated as a left R-module.
Let A be a left ideal of R[x].
If A = {0}, then clearly it is finitely generated left R[x] module.
Suppose A {0}. Let
Ik = { a R : a is leading co-efficient of some polynomial of degree k in A} {0}.
Here k 0 . We prove that Ik is left ideal of R.
Clearly, I as 0 I k .
Let a, b I k . If a = b, then a - b = 0 I k .
Suppose a b a b 0.
114 Noetherian and Artinian Modules
Then either a 0 or b 0 .
If a 0 , then there exist f(x) = axk + a1xk-1 +... A . Now, a 0 a 0 .
Now f (x) A and A is left ideal.
f (x) A a Ik .
Similarly, if b 0 b I k .
(i) if b 0 , then a b a I k .
(ii) if a 0 , then a b b I k .
(iii) if a 0 , b 0 and a b 0 . Then,
f(x) - g(x) = (a-b)xk + (a1-b1)xk-1 +... A ,
where g(x) = bxk + b1xk-1 +... A
a b Ik .
Hence, Ik is a subgroup of R under addition.
Let r R and a I k .
If ra 0 ., then ra I k .
If ra 0 , then a 0 . Now, r R R[ x ] and f ( x) A
rf ( x ) A.
Now, rf ( x) rax k ra1 x k 1 ... is of degree k so ra I k .
Hence, Ik is a left ideal of R.
If a I k and a 0 , then
xf ( x) axk 1 a1 xk ... A is of degree k + 1.
a I k 1.
Hence, I k I k 1 k 0.
Now, we have an ascending chain I 0 I1 I 2 ... I k 1 ... of left-ideals of R.
Since R is left Noetherian, therefore there exists a positive integer d such that
I d I d i i 0.
Abstract Algebra 115
Since R is left Noetherian, therefore each ideal of R is finitely generated as a left R-module. Let
nk
Ik ak1 , ak2 ,..., akn , that is, I k Rak j .
k
j 1
Now, by definition of Ik; there exists polynomial
f k j ( x ) ak j x k a1k j x k 1 ..., where 1 j n k .
d , nk d nk
Let T f k j ( x) f k j ( x) .
k 0, j 1 k 0 j 1
Let B T (The left ideal generated by T)
d nk
B R[ x] f k j ( x).
k 0 j 1
We shall prove that A = B.
Let f ( x) A and suppose deg(f(x)) = n. We write
f(x) = axn + a1xn-1 +...
We prove the result by induction on n.
If n = 0, then f(x) = ax0 (a 0) and a I0 .
n0
I 0 a01 , a02 ,..., a0 n0 Ra0 j
j 1
a r1a01 r2 a02 ... rn0 a0 n0 , rj R
n0
rj f 0 j ( x )
j 1
aB ,
that is, f ( x) B
Let n 1 and suppose that the result holds for all polynomials in A. that is, g ( x ) A and deg(g(x)) < n.
g ( x ) B.
Now either n d or n < d.
Case-I. Suppose n d and a I n I d and
nd
I d Radj [ k d ]
j 1
116 Noetherian and Artinian Modules
nd
I d rj adj .
j 1
Consider the polynomial
nd
g ( x) rj x nd f dj ( x) B.
j 1
Now, f dj ( x) is a polynomial of degree d with leading co-efficient rj adj .
nd
Hence, the leading co-efficient of g(x) is r a
j 1
j dj a and it is a polynomial of degree n.
Now g ( x ) B and B A
f ( x) g ( x) A and it is a polynomial of degree n - 1 and hence by induction hypothesis
f ( x ) g ( x ) B.
Since g ( x ) B , it follows that f( x) B. This completes the induction.
Case-II. n < d.
nm
Now, I n Ranj and a I n
j 1
nm
a rj anj , rj R.
j 1
nm
Consider g ( x) rj f nj ( x) B.
j 1
nm
Now, g(x) is of degree n and leading co-efficient is r a
j 1
j nj a.
Hence f ( x ) g ( x) A is of degree atmost n - 1 and hence by induction hypothesis f ( x ) g ( x ) B.
Since g ( x ) B f ( x ) B.
Hence, in either case A B.
But B A.
A B.
Since B is finitely generated, therefore A is finitely generated.
Abstract Algebra 117
7.3.1. Proposition. (Converse to Hilbert Basis Theorem) If R[x] is left Notherian, then so is R.
Proof. Let I be the ideal of R[x] generated by x over R[x].
that is, I = <x> = R[x]x.
Then I is an ideal of R[x].
Define a map f : R R[ x] by
I
f(a) = a + I.
Now, f(a + b) = a + b + I = (a + I) + (b + I) = f(a) + f(b)
and f(ab) = ab + I = (a + I)(b + I) = f(a)f(b).
Thus, f is ring homomorphism.
Now, let f(a) = f(b)
a + I = b + I
a - b + I = I = R[x]x
a - b ∈ I = R[x]x
a - b = f(x)x for some f(x) ∈ R[x]
f(x)/ a - b.
If a b , then deg(a - b) = deg(f(x)) + 1.
0 = deg(f(x)) + 1, a contradiction.
Thus, a = b.
Hence f is one-one.
Let g ( x) I R[ x] , g ( x) R[ x].
I
Suppose, g ( x) b0 b1 x ... bn x n
b0 (b1 b2 x ... bn x n1 ) x
b0 h( x) x
g ( x) I b0 h( x) x I b0 I . [ h(x) R[x] h(x) x I]
Now, b0 R and f (b0 ) b0 I g ( x) I .
Hence, f is onto.
Therefore, f is an isomorphism of ring R. that is, R R[ x] .
I
Since R[x] is left Noetherian and a factor ring of Noetherian is again Noetherian. It follows that R is
Noetherian.
118 Noetherian and Artinian Modules
7.3.2. Theorem. Let R be a left Noetherian ring and x1 , x2 ,..., xn be n independent indeterminants. Then,
R[ x1 , x2 ,..., xn ] is also Noetherian.
Proof. We know that R is left Noetherian iff polynomial ring R[x] is Noetherian.
We prove the result by induction on n.
If n = 1, then, R[x1] is left Notherian.
If n = 2, then, R[ x1 , x2 ] R[ x1 ][ x2 ] , is left Noetherian, since R[x1] is Noetherian.
Suppose that n > 2 and R[ x1 , x2 ,..., xn1 ] is left Noetherian.
Then, R[ x1 , x2 ,..., xn ] R[ x1 , x2 ,..., xn1 ][ xn ] is left Noetherian.
This completes the proof.
7.3.3. Theorem. Let S be a sub-ring of R such that 1 R 1 S . Suppose S is left Noetherian and R is
generated by finite number of elements as an algebra over S. Then, R is also left Noetherian.
Proof. Let x1 , x2 ,..., xn R and R is generated by x1 , x2 ,..., xn as an algebra over S, that is, every element
of R is a polynomial function in the variables x1 , x2 ,..., xn with co-efficient in S, that is, R[ x1 , x2 ,..., xn ] .
Since S is left Noetherian, by above theorem, S[ x1 , x2 ,..., xn ] is left Noetherian and so is R.
7.4. Nil Left Ideal.
Let R be a ring and I be a left ideal. If given a I , there exists a positive integer n depending upon 'a'
such that a n 0 , then I is called nil left ideal of R.
If a R and a n 0 for some positive integer n, then a is called nilpotent element of R. Thus, a left ideal
I of R is nil if every element of I is nilpotent.
7.4.1. Nilpotent Ideal.
A left ideal I of R is called nilpotent if there exists a positive integer n such that In = {0}. that is, for
every a1, a2,...,an ∈ I,
a1.a2...an = 0.
In particular, for every a∈I,
.a...a 0 , that is, an = 0
a
n times
that is, a is nilpotent element of R. Hence, a nilpotent left ideal is a nil left ideal.
7.4.2. Theorem. In a left Artinian ring, every nil left ideal is nilpotent.
Proof. Let R be a left Artinian ring and I be a nil left ideal of R.
We shall prove that I is nilpotent.
Abstract Algebra 119
Consider the decreasing chain I I 2 I 3 ... I m I m 1 ... of left ideals of R.
Since R is left Artinian, there exists a positive integer k such that Ik = Ik+1 =...
that is, I k I k i i 0.
Let J = Ik. Then, J is left ideal of R such that J2 = [Link] = Ik = J.
We claim that J = {0}.
Suppose J {0}.
Let £ = {A : A is left ideal of R contained in J such that JA {0}}
Now, JJ = J2 = J {0} and J is left ideal of R contained in J.
J £ £ .
We know that if R is left Artinian, then every non-empty set of left ideals of R has a minimal element.
Let A be the minimal element of £ such that JA {0}.
Hence, there exists a( 0) A such that Ja {0}, otherwise JA = {0}.
Also, a A and A is left ideal of R contained in J.
B Ja A J .
Thus, B is a left ideal of R and
JB = JJa = Ja = B {0}.
B £.
Since B A and A is the minimal element of £ .
B A.
that is, A = Ja = B.
Now, a A Ja .
a xa for some x J .
Now, a = x(xa) = x2a, a simple induction shows that a = xna n 1 .
Now, x J I implies that x is a nilpotent element.
Let m be a positive integer such that xm = 0.
Then, a = xma = 0a = 0, that is, a = 0, a contradiction.
Thus, J = 0, that is, Ik = {0}, that is, I is a nilpotent left ideal of R.
Note. A ring with unity is called a division ring if every non-zero element of R is invertible.
120 Noetherian and Artinian Modules
7.4.3. Proposition. An Artinian domain is a division ring (Skew-field).
Proof. To prove that R is a division ring, we shall prove that every non-zero element of R is invertible.
Let a ∈ R be a non-zero element.
Let I k a k Ra k , that is, left ideal of R is generated by ak.
Now, if y ∈ Ik+1 y = rak+1 = (ra)ak ∈ Ik, r ∈ R, thus I k 1 I k k 0.
So, we have a decreasing chain
I1 I 2 ... I n I n1 ...
of left ideals of R.
Since R is left Artinian, so there exists a positive integer k such that
I k I k i i 0
that is, a k a k i i 0 .
Now a k a k a k 1
ak ra k 1 for some r R
ak raak
(1 ra)ak 0.
Now, R is a division ring and a 0 a k 0 and therefore 1 ra 0 , that is, ra 1 , that is, every
element of R is left invertible. By the same reasoning, there exists b R such that 1 br.
Then, a (br ) a b( ra ) b.1 b.
Hence, r is the inverse of a and hence R is a division ring.
7.4.4. Prime Ideal. Let R be a ring. An ideal P of R is called a prime ideal if given ideals A and B such
that AB P implies either A P or B P .
Remark. In a left Artinian ring every prime ideal is maximal.
7.4.5. Theorem. Let R be a left Noetherian ring. Every ideal I of R contains a finite product of prime
ideals.
Proof. Let R be a left Noetherian ring. We shall prove that every ideal of R contains a finite product of
prime ideals. Let
£ = {I : I is ideal of R such that I does not contain a finite product of prime ideals}.
To prove the result, we prove that £ = ϕ.
Suppose £ .
Abstract Algebra 121
Since every ideal is a left ideal and R is left Noetherian, therefore £ must have a maximal element, say I.
Now, I £, so I is not prime ideal, so there exist ideals A and B of R such that the product AB I
implies neither A I nor B I .
Consider the ideals, I1 = I + A and I2 = I + B.
Since I is maximal ideal of £, therefore I1 and I2 £.
Hence there exist finite number of prime ideals, say P1 , P2 ,..., Pm and P1 , P2 ,..., Pn such that
1 2 ...Pm I1 and P1 P2 ...Pn I 2 .
PP
Now, P1 P2 ...Pm P1 P2 ...Pn I1 I 2
Let x I1I 2 , then x yz,
finite
i i yi I1 , zi I 2 .
Now, yi I1 I A yi ci ai where ci I , ai A
and zi I 2 I B zi di bi where di I , bi B .
Then, yi zi (ci ai )(di bi ) ci di ci bi ai di ai bi I [ ci ,di I , ai A, bi B & AB I ]
Hence, x y z I .
finite
i i
From this, it follows that P1 P2 ...Pm P1 P2 ...Pn I , that is, I contains a finite product of prime ideals, a
contradiction and this contradiction proves that £ = ϕ, that is, every ideal of R contains a finite product
of prime ideals.
7.4.6. Boolean Ring. A ring R is called Boolean Ring if x2 x x R. Suppose R is a Boolean ring,
then x2 x x R. Then
( x y)2 ( x y)
x 2 y 2 xy yx x y
x y xy yx x y
xy yx x, y R .
Also, ( x x)2 ( x x)
xx 0
x x x R .
Now, xy xy ( yx) yx .
Therefore, R is a commutative ring.
122 Noetherian and Artinian Modules
Remark. Let R be a Boolean Noetherian ring, then
RZ Z ... Z
2Z 2Z 2Z
and so if a Boolean ring which is also Noetherian, then number of elements in R is 2n where n is a
positive integer.
Remark. If a, b R (ring) such that ab = 1 and ba 1. Define eij b i 1a j 1 bi a j .
0 if j k
We shall prove that (i) eij ekl
eil if j k
(ii) eii 0
(iii) eii 1.
We have, eij ekl (b i 1a j 1 b i a j )(b k 1a l 1 b k a l )
b i 1a j 1b k 1a l 1 bi a j b k 1a l 1 bi 1a j 1b k a l b i a j b k a l .
Suppose j k . Then, either j k or j k .
First suppose that j k .
j 1 k 1.
Now, eij ekl b i 1a j k a k 1b k 1a l 1 b i a j k 1a k 1b k 1a l 1 b i 1a j 1 k a k b k a l b i a j k a k b k a l
Since a mbm 1 m 0.
eij ekl b i 1a j k a l 1 b i a j k 1a l 1 bi 1a j 1 k a l b i a j k 1a l 1
b i 1a j l k 1 b i a j l k bi 1a j l 1 k b i a j l k 0.
Similarly, we can prove that eij ekl 0 if j k .
Suppose j k , then
eij ekl eij e jl b i 1a j 1b j 1a l 1 b i a j b j 1a l 1 b i 1a j 1b j a l b i a j b j a l
bi 1a l 1 bi a l bi a l b i a l bi 1a l 1 bi a l eil .
0 if j k
Hence, ekl .
eil if j k
(ii) Take fi eii .
Then, fi fi eii eii eii fi .
Abstract Algebra 123
fi is idempotent.
Suppose fi 0 , then
f i eii bi 1a i 1 bi a i 0
bi 1ai 1 bi ai
ai 1bi 1ai 1bi 1 ai 1bi aibi 1
1 ba, a contradiction.
Hence eii 0.
(iii) Suppose fi 1. Take j i , then
f i f j eii f j f j . ( eii 1)
Again, f i f j eii e jj 0. [ i j ]
Hence f j 0 , a contradiction.
Hence, eii 1.
7.4.7. Theorem. Let R be left Noetherian ring and a, b R such that ab 1 , then ba 1.
Proof. Let R be a left Noetherian ring and a, b R such that ab 1 . We have to show that ba 1.
Suppose that ba 1 . Now, ab 1
a (ab) b ab
a 2b2 1.
A simple induction shows that a mbm 1 m 0.
Define, eij b i 1a j 1 bi a j . (i 1, j 1)
Now, as shown in above remark
0 if j k
eij ekl ,
eil if j k
eii 0 and eii 1.
Define, fi eii and let I k Rf1 Rf 2 ... Rf k .
Then, Ik is a left ideal of R and is contained in Ik+1 k 1.
124 Noetherian and Artinian Modules
that is, I k I k 1 k 1.
Now, f k 1 Rf k 1 Rf1 Rf 2 ... Rf k 1 Ik 1 .
Suppose, f k 1 Ik . Then, f k 1 r1 f1 r2 f 2 ... rk f k where ri R
f k 1 f k 1 r1 f1 f k 1 r2 f 2 f k 1 ... rk f k f k 1 0 [ f r f s 0 if r s]
f k 1 0 , a contradiction. [ f r is idempotent]
Hence I k I k 1 .
Hence, we get an infinite properly ascending chain
I1 I 2 I 3 ... I k I k 1 ...
of left ideals of R, a contradiction because R is left Noetherian.
Hence, ba = 1.
7.4.8. Theorem. Let R be a Noetherian ring having no non-zero nilpotent ideals. Then R has no non-
zero nil ideals.
Proof. Let, if possible, A be a non-zero nil ideal of R.
Let = { l (a) : a A, a 0} be a family of left annihilators of all non-zero elements of A. Since
each l (a) is a left ideal of R and A {0}, so is a non-empty family of left ideals of R. Since R is a
Noetherian ring, so must have a maximal element, say, l (a).
Now, let x R be any arbitrary element, then ax A [ A is an ideal of R]
But A is a nil ideal so ax must be a nilpotent element that is, there exist a smallest positive integer k such
that
(ax)k = 0 (ax) (ax)k–1 = 0 ax l (ax) k --1 (1)
Now, we see that (ax)k–1 0 and (ax)k–1 A l (ax) k --1
Now, we prove that l(a) l ( ax ) k --1
Let y l(a) ya = 0 (ya)x = 0 y(ax) = 0
y(ax)k–1 = 0 y l (ax) k --1 l(a) l (ax) k --1
But l(a) is a maximal element of , so l(a) = l (ax) k --1 (2)
By (1) and (2), we have ax l(a)
axa=0 x R
aRa=0
Abstract Algebra 125
Let J = R a R, then, J2 = R a R R a R R a R a R [ R R R]
= R (a R a) R = {0}
that is,, J is a nilpotent ideal of R.
But R has no non-zero nilpotent ideal, so we must have
J = {0} that is,, RaR = {0} (3)
Now, consider the ideal
B=<a>=Ra+aR+RaR+a
Let D=Ra+aR+RaR=Ra+aR [By-(3)]
Then, D2 = (Ra + aR) (Ra + aR) = R a R a + a R R a + R a a R + a R a R
R a R a + a R a + R a R + a R a R [ R2 R, RaaR RaR] = {0}
as R a R = a Ra = 0, that is, D is a nilpotent ideal of R. But R has no non-zero nilpotent ideal, so we
must have D = {0}
Hence, we obtain B = a
Now, a A and A is a nil ideal so ‘a’ must be a nilpotent element that is, there exists a smallest
positive integer t such that at = 0.
Then, Bt = a .a ...a (t times) = at = {0}, that is, B is nilpotent ideal of R. But R has no non-
zero nilpotent ideal, so B = {0}
a = {0}
Now a.1 a = {0}
a.1 = 0
a = 0, a contradiction.
Hence R has no non-zero nil ideal.
7.4.9. Proposition. Let R be a left Noetherian ring. Prove that sum of nilpotent ideals of R is again a
nilpotent ideal of R.
Proof. Let R be a left Noetherian ring and I be a family of nilpotent ideals of R.
Suppose, I I .
Then, I is an ideal of R, since sum of ideals is again an ideal.
Since R is left Noetherian, therefore I is generated by finite number of elements, say x1 , x2 ,..., xn I .
Now, xi I there exists a finite subsets i of such that xi I .
i
126 Noetherian and Artinian Modules
n
Take, i . Then, xi
i 1
I
i
I x1 , x2 ,..., xn I I
I I
and is a finite subset of . Thus, we may assume that I I1 I 2 I3 ... I m and each Ik is
nilpotent.
Thus, to prove the result it is equivalent to prove that sum of finite number of nilpotent ideals is again a
nilpotent ideal.
We prove the result by induction on m.
If m = 2, then
I1 I 2 I1
.
I2 I1 I 2
I1 I1 I 2
Now, I1 is nilpotent ideal of R, so is nilpotent ideal of R . Hence, is a nilpotent
I1 I 2 I1 I 2 I2
ideal. since I2 is a nilpotent ideal of R, it follows that I1+I2 is a nilpotent ideal of R.
Thus, the sum of two nilpotent ideals is again a nilpotent ideal of R.
Let m > 2 and suppose that the result hold for (m - 1) nilpotent ideals, that is, I1 I 2 I3 ... I m1 is
nilpotent.
Now, sum of two nilpotent ideals is again nilpotent, so the sum of I1 I 2 I3 ... I m1 and I m , that is,
I1 I 2 I3 ... I m is also nilpotent.
Remark. Let R be a ring and A and B be ideals of R such that A B . Then, B is nilpotent iff both A
and B are nilpotent.
A
7.4.10. Proposition. In a left Noetherian ring, every nil ideal is nilpotent.
Proof. Let £= I be the family of all nilpotent ideals of R and I I .
Then, I is an ideal of R. Also, we know that in a left Noetherian ring sum of nilpotent ideals is again
nilpotent and hence I is nilpotent ideal of R.
Consider the quotient ring R .
I
Let J be the nilpotent ideal of R , where J is an ideal of R containing I. Now, J is an nilpotent
I I I
ideal of R and I is nilpotent ideal of R. Hence, J £ .
I
Thus, there exists 0 such that J I 0 .
Abstract Algebra 127
Now, I I I 0 J I
I J.
Hence, J is the zero ideal. that is, J = {I}.
I I
Therefore, R has no non-zero nilpotent ideal. Since R is left nilpotent, therefore, R is left
I I
Noetherian.
Thus, R is left Noetherian ring such that R has no non-zero nil ideal.
I I
Let k be a nil ideal of R, then K + I is an ideal of R
KI is an ideal of R .
I I
Now, K I K K I is nil ideal. Hence, K I is a nil ideal of R .
I K I and K is nil ideal, so I I I
Since K I has no non-zero nil ideal.
I
Therefore, K I {I } K I I K I.
I
Since I is a nilpotent ideal, it follows that K is also a nilpotent ideal of R. Thus, every nil ideal in R is
nilpotent.
7.4.11. Example. Example of a ring which is both Noetherian as well as Artinian.
Proof. Let R be a ring having only finite number of left ideals. Clearly, every properly ascending chain
or a properly descending chain is finite and therefore R is left Noetherian as well as left Artinian.
7.4.12. Example. Example of a ring which is Noetherian but not Artinian.
Solution. Let R=Z, the ring of integers. Now, since Z is a P.I.D., so every ideal of R is generated by a
single element that is, if I is an ideal of R, then ∃ m ∈ I, m 0 such that I=mZ. that is, every ideal of R is
finitely generated, therefore R is Noetherian.
Let Ik = 2k Z, k 0
Now, 2k+1∈ Ik+1, and 2k+1 =2. 2k ∈ 2k Z= 2k =Ik, so 2k+1 ⊆ Ik and thus Ik+1⊆ Ik
Now, 2k ∈ Ik.. If 2k ∈ Ik+1, then 2k = 2k+1z, where z ∈Z, so 1=2z , ehich implies
z = ∈ Z, a contradiction.
Hence 2 k I k 1 and so Ik+1 Ik
Thus,
I 0 I1 ... I n ...
is an infinite properly descending chain of ideals of R.
Therefore, R is not Artinian.
128 Noetherian and Artinian Modules
7.4.13. Example. Example of a ring which is neither Noetherian nor Artinian.
Solution. Let R be a ring and x1, x2, ..., xn... be infinite numbers of commutingindeterminants over R.
Consider the polynomial ring R[x1, x2, ..., xn]=S. Let Ik be the left ideal of S generated by x1, x2, ..., xn,
that is,
Ik = x1, x2, ..., xn = Sx1 +Sx2+...+SxK.
Now, it is clear that Ik Ik+1
If xk+1∈ Ik+[Link],
xk+1= r1x1+ r2x2+...+ rkxk where ri∈ S = R[x1, x2, ..., xn...]
The L.H.S. is a polynomial in the single indeterminates xk+1 while the R.H.S. is a polynomial in the
indeterminants x1, x2, ..., xn.
Therefore these two polynomials cannot be equal.
Hence, xk+1 ∉ Ik ⇨ Ik⊆ Ik+1. Thus, we have
I1 I 2 ... I n ...
is an infinite properly ascending chain of left ideals of S
Hence,S is not left Noetherian.
Similarly, we can prove that S is not right Noetherian.
Let, B0={x1,x2,... xn,...}
B1= B0-{x1}
B2= B0-{x1,x2}
. . .
Bn =B0-{x1,x2,... xn}
Let Jn be the left ideal of R[x1,x2,... xn,...]=S generated by Bn.
Now, it is clear that
J 0 J1 ... J n ...
is an infinite properly descending chain of left ideals of S.
Therefore, S is not left Artinian.
Similarly, we can prove that S is not right Artinian.
7.4.14. Proposition. Let a P.I.D and I 0 be an ideal of R. Then, R/I is both Noetherian and
Artinian.
Proof. Since R is P.I.D.,so every ideal of R is generated by a single element. Thus, every ideal of R is
finitely generated and so R is Noetherian.
Abstract Algebra 129
Now, R is Noetherian and a quotient ring of a Noetherian ring is a Noetherian, so R/I is Noetherian.
Let a ∈I be such that a= aR= . Since I {0}, so a 0.
If a=1 or a unit, then I=R and
R/I= R/R ≅ {R}= the zero ring .
But zero ring is trivially Artinian ring so R/I is Artinian.
Suppose, a is not a unit, then a a11 a2 2 ...am m where each ai i is an irreducible element of R and each i´
s are positive integers.
Let J/I be an ideal of R/I. Here, J is an ideal of R, so there exists an element b∈ J such that
J = bR = < b>
I = < a >⊆ J = <b>
which implies a= br for some r∈ R .
So, b/a.
Thus, b a11 a2 2 ...am m where 0
Now each can be selected in ( +1) ways.
Therefore, the number of choices for b is ( +1) ( +1)... ( +1), so the number of choices for b is
( +1) ( +1)... ( +1), which means the number of choice for J/I is
( +1) ( +1)... ( +1)
Thus, R/I has only finite number of ideals.
We know that a ring which has finite number of ideals is an Artinian ring.
7.5. Ring of Homomorphisms. The collection of all homomorphisms from a R-module M to itself is
denoted by HomR M , M and is a ring.
7.5.1. Opposite ring. Let (R, +, . ) be a ring. Then the opposite ring of R, denoted by Rop, is defined as
the ring (R,+, o), where the operation o is given as
x y y . x for all x, y R
Results.
(i) Let R be a ring and Rn denote the ring of n n matrices over R then (R n )op (R op ) n
(ii) If R is a division ring then R op is also a division ring.
(iii) If a ring R is direct sum of rings R1, R2, …, Rk that is,, R = R1 R2 ... Rk, then
R op = R 1op R op
2 ... R k .
op
130 Noetherian and Artinian Modules
(iv) Let M be a R-module such that M = M
is the sum of a family of simple R-modules M ,
then there exists a sub- family M such that M = M
In words, we can say that if a R-module M is sum of simple R-modules, then it can be represented as the
direct sum of a sub family of family of these simple R-modules.
Now, we will state some lemma´s, which will be useful for proving the Wedderburn- Artin Theorem.
7.5.2. Lemma. Let M M1 M 2 ... M n be a left R-module (Mi ´s are submodules of M) and let Aij
= HomR (Mi Mj ). Then,
A11 A12 ... A1n
A A22 ... A2 n
HomR (M, M)= 21 .
... ... ... ...
An1 An 2 ... Ann
7.5.3. Lemma. Let M M1 M 2 ... M n , where Mi ≇Mj, i≠j and each Mi is a simple R-sub module
of M. Then
D1 0 0 ... 0
0 D2 0 ... 0
HomR (M, M)=
... ... ... ... ...
0 0 0 ... Dn
where each Di=Aii and is a division ring.
7.5.4. Lemma. Let M M1 M 2 ... M n , where Mi is simple and M i M j ∀ i and j. Suppose
HomR M 1 , M 1 End R M 1 D then, HomR (M, M) ≅ Mn(D)
D D ... D
D D ... D
where, Mn(D) = , where D is a division ring.
... ... ... ...
D D ... D
7.5.5. Lemma. Let A be minimal left ideal of R. Then either A2={0} or A=Re, e is an idempotent in A.
7.5.6. Wedderburn- Artin Theorem. Let R be left (or right) artinian ring with unity and no nonzero
nilpotent ideal. Then, R is isomorphic to a finite direct sum of matrix rings over division rings.
Proof. First, we prove the following lemmas:
[Link]. Lemma. Let R be a left Artinian ring with unity having no nonzero nilpotent ideals. Prove that
every nonzero left ideal of R contains nonzero idempotents.
Abstract Algebra 131
Proof. Let A be a nonzero left ideal of R and let
J1={ B⊆A : B is a nonzero left ideal of R}
Now, A≠ {0}and A⊆A ⇨A∈ J1, then
J1 ≠ ∅
Since R is left Artinian, therefore, J1 has a minimal element, say B.
Then, B is a minimal left ideal of R.
Then, either B2 = {0} or B = Re for some idempotent e ∈ B.
Suppose B = {0}.
Consider the ideal J= BR
Then, J2 = BRBR ⊆ BBR = B2R= {0}.
Thus, J is a nilpotent ideal of R. By the given condition J = {0}, that is, BR = {0}.
Now, b = b.1∊ bR⊆ BR = {0}, which implies b=0 for all b∈ B and so B = {0}, a contradiction. Hence,
B= Re
Now, B⊆ A and e = e.1 ∊ Re = B ⊆A, so
e ∊ A.
Thus, A contains nonzero idempotents.
[Link]. Lemma. Let R be a left artinian ring with unity having no nonzero nilpotent ideals. Then, every
left ideal of R is generated by an idempotent.
Proof. Let A be any left ideal of R and let S be the set of all nonzero idempotent of A. Then, S ≠ ∅,
since by lemma [Link]., there exists an idempotent e ∊ A.
Now, consider the left ideal R(1-e) ∩ A, e ∈ S.
and let
J = { R(1-e) ∩ A, e ∈ S }
Since S is nonempty, so J ≠∅
Again R is left Artinian, since J has a maximal element, say R(1-e) ∩ A where e ∈ S.
We claim that R(1-e) ∩ A= {0}
Suppose R(1-e) ∩ A ≠ {0}. Then, it is a nonzero left ideal of R. Therefore, by lemma [Link]., this left
ideal has a nonzero idempotent, say e1.
Now, e1 ∈ R(1-e) and e1 ∈A
Now, e1 ∈A e1 ≠ 0 ⇨ e1 ∊ S.
Also, e1∈ R(1-e) ⇨ e1 = r(1-e) for some r∊ R, so
132 Noetherian and Artinian Modules
e1 e = r(1-e)e = r(e-e2 ) =r(e-e) =r.0 =0
Let e' =e+ e1- e e1 ∊ A. Then
e′ = e' e' =( e+ e1- e e1 )( e+ e1- e e1 )
= e2+ e e1 - e2 e1 + e e1 + e12- e1 e e1 - e e1 e- e e12 + e e1 e e1
= e+ e e1 - e e1 +0 + e1- 0.e1 -e.0- e e1 + e .0.e1
= e+ e1 - e e1 = e'
and e1 e' = e1e + e12 - e1ee1 = 0+e1 - 0. e1
= e1 ≠ 0
Thus, e1 e' ≠ 0 ⇨ e' ≠ 0 ( *)
Hence,e' is an idempotent and so e'∊ S.
But then R(1-e' ) ∩A ∊ J
Now, r1(1-e')= r1(1-e -e1 + e e1 ) = r1(1-e)- r1r(1-e)+ r1e r(1-e) = R(1-e) [ r, r1, e ∊ R)
So, R(1-e')⊆ R(1-e)
and thus, R(1-e' ) ∩A ⊆ R(1-e) ∩A.
Now, e1∊ R(1-e) ∩A.
Suppose, e1∊ R(1-e' ) ∩A. Then, e1 = s(1-e' ), s∊ R, therefore
e1 e' = s(e' - e'2) = s(e' - e') =0
a contradiction, because e1e' ≠ 0 ( by *)
Hence, e1 ∉ R(1 - e' ) ∩A
However, in that case R(1 - e' ) ∩A R(1 - e) ∩A.
Now, R(1 - e' )∩A and R(1 - e)∩A ∊ J2 and R(1 - e) ∩ A is the minimal element of J2, a contradiction.
The contradiction proves that
R(1 - e) ∩A = {0} (**)
Let a∊ A, then a(1 - e) = a - ae ∊ A and a(1 - e) ∊ R(1 - e)
implies, a(1 - e) ∊ R(1 - e) ∩A = {0}
implies, a(1 - e) ∊ {0}
implies, a(1-e) =0 for all a ∊ A
implies, a = ae for all a ∊ A
implies, A⊆ Ae ⊆ Re ⊆A
implies, A = Re
Abstract Algebra 133
Hence, every left ideal of R is generated by an idempotent.
[Link]. Lemma. Let R be a left Artinian ring with unity having no non-zero nilpotent ideal. Let J be the
set of all non -zero minimal left ideals of R. Prove that R= I ; J ={ I⋋:⋋∈∧}.
Proof. Let J be the set of all minimal left ideals of R.
Since R is left Artinian, so J ≠φ. Let J = { I⋋ }⋋∈∧. Put I= I . Then, I is a left ideal of R. By lemma
[Link]., there exists an idempotent e ∊ I such that I = Re.
We claim that R(1-e) ={0}. Now, this left ideal will contain a minimal left ideal of R, say B, so
B ∊ J, which implies B ⊆ I = Re
Now, B ⊆ Re ∩ R(1 - e) = {0} (By (**) )
which implies B = {0}
A contradiction. Hence,
R(1-e) ={0} ⇨R =Re.
Hence, I = Re = R, and so R= I .
Proof of the theorem . We know that a minimal left ideal is a simple submodule of RR.
⇨ RR is the sum of simple submodules.
Therefore, there exists a subset ∧' ⊆∧ such that
RR = ⊕∑I⋋.
Now, 1∈R = RR, therefore
1 = x⋋1+ x⋋2+...+ x⋋n, where ⋋i∈∧'
so, r = rx⋋1+r x⋋2+...+r x⋋n,
so, R ⊆ I⋋1+ I⋋2+...+ I⋋n ⊆ RR
so, RR ⊆ I⋋1+ I⋋2+...+ I⋋n
We may assume that
R = RR = I1⊕I2 ⊕ ... ⊕ In.
where each Ij is a minimal left ideal of R.
Suppose Ij = Rej where ej ϵ Ij is an idempotent.
Thus we may write
R = RR = Re1 ⊕ Re2 ⊕ ... ⊕ Ren
134 Noetherian and Artinian Modules
Suppose
R = RR = (Re1 ⊕ Re2 ⊕ ... ⊕ Ren) ⊕ (Re1 ⊕ Re2 ⊕ ... ⊕ Ren)
where nk =n, where every two ideals in every bracket are isomorphic as left R-module and the ideals in
the different brackets are not isomorphic. ( This can be done if necessary by rearranging and
renumbering the ideals in the desired form).
Take R = M1 ⊕ M2 ⊕ ... ⊕ Mk, where Mi is the sum of left ideals in the ith bracket.
A11 A12 ... A1k
R R
Then, Hom ( R , R ) ≅ ... ... ... ...
A
k1 Ak 2 ... Akk nn
where Aij = HomR (Mi, Mj )
We know that
(i) if X = X1 ⊕ X2 ⊕ ... ⊕ Xm, where each Xi is a left R- module of M such that
HomR (Xj, Xi )={0} for all i ≠ j.
A11 0 ... 0
0 A22 ... 0
Then, HomR (Xi, Xj ) = , where Aij = HomR (Xj, Xi ).
... ... ... ...
0 0 ... Amm
(ii) If B = B1 ⊕ B2 ⊕ ... ⊕Bm and C = C1 ⊕ C2 ⊕ ... ⊕Ct, where Bi's & Ci's are simple
submodules of M & Bi≇ Cj for all i ≠ j, then HomR (B, C ) = {0}
(iii) if A= A1 ⊕ A2 ⊕ ... ⊕An and each Ai is simple and Ai ≅ Aj for all i ≠ j, then,
HomR (A, A ) ≅ Mm (D), where D = HomR (A1, A1 ) is a division ring.
Using all these, we conclude that
Aij = HomR (Mj, Mi ) = {0} if i≠j.
and Aii = HomR (Mi, Mi ) ≅ M ti Di ,
i 1
i 1
where Di≅ HomR Ren 1 , Ren 1 and Mi= Re ni1 1 ... Re ni
Hence,
M t 1 (D1 ) 0 ... 0
0 M (D ) ... 0
Hom R R , R R
t 2 2
... ... ... ...
0 0 ... M t k (Dk )
M t1 D1 M t2 D2 ......... M tk Dk
Abstract Algebra 135
We Know that Rop ≅ HomR (RR, RR)
Hence, Rop is isomorphic to finite direct product of matrix rings over the division rings .
From this we conclude that R = ( Rop)op is also isomorphic to finite direct product of matrix rings over
the division ring.
7.5.7. Maschke Theorem. If F is the field of complex numbers and G is a finite group, then
F G Fn1 ... Fnk
for some positive integers n1 ,..., nk .
Proof. We first prove that F G has no non zero nilpotent ideals.
Let G g1 e, g 2 ,..., g n , and x i g i F G . Set x* i g i1 , where i denotes the complex
conjugate of i . Then
n n
xx* i i g i
2
i 1 i2
2
for some i F . Hence, xx* 0 implies i 0 , so each i 0 ; that is x 0 . Thus, xx* 0
i 1
implies x 0 . Let A be a nilpotent ideal in F G . Let a A . Then aa* A , so aa * is nilpotent, say
r
aa * 0 . (We may assume r is even.) Set b aa * 2 . Then b 2 0 and b b * . Thus, bb* 0 , which
r
r
gives aa * 2 b 0 . Proceeding like this, we get aa* 0 . Hence, a 0 , which proves that A 0 .
Hence, F G has no nonzero nilpotent ideals.
Further, F G is a finite-dimensional algebra with unity over the field F. Therefore, F G is an
artinian ring. Then by the Wedderburn –Artin theorem,
F G Dn1 ... Dnk ,
1 k
Where D i ,1 i k , are division rings. Now each Dnii contains a copy K of F in its center. In this way
each Dnii is a finite dimensional algebra over K (How?). Let D i : K n , and a D . Then
i
1, a, a2 ,..., an are linearly dependent over K. Thus, there exist 0 , 1 ,..., n (not all zero) in K such that
0 1a ... n a n 0 . But since K is algebraically closed, 0 1 x ... n x n K x has all its roots
in K. Hence, a K , which shows that D i K F and completes the proof.
136 Noetherian and Artinian Modules
7.6. Radical ideal. A two-sided ideal I in a ring R with unity is called a radical ideal with respect to a
specified property P if
1 . the ideal I possesses the property P and
2. the ideal I is maximal for the property P, that is, if J is a 2-sided ideal of R having the property P, then
J I.
7.6.1. Jacobson radical. Let R be a ring with unity, then the Jacobson radical denoted by J(R) is a
radical ideal of R defined as
J(R) = {a R : 1 - a is a unit in R}.
7.6.2. Example. In the ring of integers, J( Z ) = ( 0 ) .
7.6.3. Radical of an Art inian Ring.
7.6.4. Proposition. The Jacobson radical of an Artinian ring is the intersection of some finitely many
maximal left (right) ideals.
Proof. Let R be an Artinian ring and be the family of all maximal left ideals of R. Let F be the family
of all left ideals of R each of which is an intersection of finitely many maximal left ideals of R.
n
Obviously F as F. Since R is Artinian, therefore, F has a minimal member, say J 0 M i ,
i 1
where M i . By definition of Jacobson radical, J(R) J0.
If M , then J0 M F and so J0 M = J0, by the minimality of J0 which means
that J0 M, for all M . Thus we get that J(R) J0 M i
= J and hence J = J0, as required.
m
7.6.5. Exercise. In a commutative Artinian ring, the only maximal ideals are M i ,1 i n , where
n
J ( R) M i .
i 1
7.7. Check Your Progress.
1. Every division ring is Noetherian.
2. An Artinian integral domain with atleast two elements is a field.
3. Subring of an Artinian ring need not be Artinian.
7.8. Summary.
In this section, we obtained results for Noetherian and Artinian. The relation between nil and nilpotent
ideals. Also, observed that every nil ideal is nilpotent but the converse statement need not be true. Also
the difference between these two is really interesting to study.
You might also like
- Understanding Left R-Modules in AlgebraNo ratings yetUnderstanding Left R-Modules in Algebra93 pages
- Algebra 4 Course Notes: Modules & FieldsNo ratings yetAlgebra 4 Course Notes: Modules & Fields85 pages
- Advanced Algebra I: Module Theory BasicsNo ratings yetAdvanced Algebra I: Module Theory Basics40 pages
- Understanding R-Modules and Their PropertiesNo ratings yetUnderstanding R-Modules and Their Properties3 pages
- Module Theory: Definitions and ExamplesNo ratings yetModule Theory: Definitions and Examples21 pages
- Mathematics Encyclopedia: Modules & Rings100% (1)Mathematics Encyclopedia: Modules & Rings1,290 pages
- Understanding Rings and Modules in AlgebraNo ratings yetUnderstanding Rings and Modules in Algebra5 pages
- Understanding Modules and Vector SpacesNo ratings yetUnderstanding Modules and Vector Spaces75 pages
- Understanding R-Modules and Their PropertiesNo ratings yetUnderstanding R-Modules and Their Properties21 pages
- Structure Theorem for Modules over PIDsNo ratings yetStructure Theorem for Modules over PIDs26 pages
- Semisimple Rings and Representation TheoryNo ratings yetSemisimple Rings and Representation Theory15 pages
- Commutative Algebra: A-Modules ExplainedNo ratings yetCommutative Algebra: A-Modules Explained21 pages
- Associative Algebras and Modules ExplainedNo ratings yetAssociative Algebras and Modules Explained19 pages
- Understanding Noetherian Rings and ModulesNo ratings yetUnderstanding Noetherian Rings and Modules60 pages
- Unitary Zn-Modules and R-Module HomomorphismsNo ratings yetUnitary Zn-Modules and R-Module Homomorphisms4 pages
- Statistical Analysis for Business InsightsNo ratings yetStatistical Analysis for Business Insights23 pages
- Locus, Lines, and Coordinate Geometry ProblemsNo ratings yetLocus, Lines, and Coordinate Geometry Problems14 pages
- S32K3xx Development Package 3.5.13 ReleaseNo ratings yetS32K3xx Development Package 3.5.13 Release8 pages
- Grammar Essentials: Conjunctions & ModalsNo ratings yetGrammar Essentials: Conjunctions & Modals62 pages
- English Unit 1: His First Flight AnalysisNo ratings yetEnglish Unit 1: His First Flight Analysis4 pages
- Asymptotic Analysis of Sorting AlgorithmsNo ratings yetAsymptotic Analysis of Sorting Algorithms51 pages