Understanding Left R-Modules in Algebra
Uploaded by
jindtannuUnderstanding Left R-Modules in Algebra
Uploaded by
jindtannuMATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 1
Def. 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 for all a, b M ; r , s R
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 ) for all a, b M ; r , s R
Remark : (1) In general , a left R – module is not a right R – module. However , if the ring R is
commutative , then a left R – module is a right R – module and vice – versa.
(2) The theory of right R – modules can be developed in same manner as the theory of left R –
modules. We shall develop here the theory of left R – modules and unless explicitly mentioned , the
term R – module will always mean “left R - module”.
Def. 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 the ring R is a field , then a unital R – module is a vector space over R. So we can say
that concept of module is a generalization of that of a vector space.
Results :
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 : By above example , we have 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
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 2
(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 .
Def. Submodule : 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
Criterion for Submodule : Let R be a ring and let M be an R – module. A non – void subset N of
M is a submodule of M iff
(i) N is additive subgroup of M
(ii) N is closed under scalar multiplication
Alternatively , A non – empty subset N of R – module M is a submodule of M iff
(i) a b N for all a, b N
(ii) ra N for all a N , r R .
Results :
1. Every R – module M has the two submodules M and {0}. these are called improper submodules.
Any other submodule of M is known as proper submodule.
2. If M is an R – module and a M then the set Ra {ra : r R} is a sub module of M.
3. 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’.
4. The intersection of two submodules of an R – module is an R – submodule.
5. An arbitrary intersection of submodules of an R – module is an R– submodule.
Def. Linear Sum of submodules : Let M be an R – module and M1 and M2 be submodule of M.
Then linear sum of M1 and M2 is denoted by M1 + M2 and is defined as :
M1 M 2 {x1 x2 : x1 M1 , x2 M 2 }
6. Linear sum of two submodules of a R – module M is a submodule of M.
7. Linear sum of finite number of submodules of a R – module M is a submodule of M.
Def. Direct sum of submodules : If A and B are two submodules of an R – module M , then M is
called the direct sum of A and B if every element ‘a’ of M can be uniquely expressed as
a a1 a2 , a1 A , a2 B . We then write M = A B
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 3
8. The necessary and sufficient condition for a module M to be a direct sum of its submodules A and B
is that (i) M = A + B (ii) A B = {0}
Def. Quotient Module : Let M be an R – module and N be a submodule of M. Then N is an additive
subgroup of M. If m M , then m + N is a coset of N in M. Now consider
M N {x N : x M}
Then M N is an R – module under the operations
( x N) ( y N) ( x y ) N
and r (m1 N) rm1 N , r R , m1 M
That M N is an R – module is proved in the following theorem. This module is known as quotient
module.
Result : Let N be an R – submodule of an R – module M , then the set
M N = {x N : x M} is an R – module for addition and scalar multiplication defined as follows :
( x N) ( y N) ( x y ) N
r ( x N) rx N for all x N , y N M N , r R .
Def. Module Homomorphism : Let M and N be R – modules. A mapping T : M N is called a
homomorphism if
(i) T(x y ) = T(x )+T(y ) , for all x, y M
(ii) T(rx) = r T(x) for all x M , r R
Results :
1. Basic properties of homomorphism : If M and N are two R – modules and if T : M N is a
homomorphism , then it is easy to check that
(i) T(0) = 0 N
(ii) T( x ) = T(x)
(iii) T(x y ) = T(x) T(y ) for all x, y M
Def. Kernel of a homomorphism : Let T : M N is a module homomorphism then the kernel
K(T) of T is defined as K(T) = {m M , T(m) = 0} , where ‘0’ is additive identity of N.
2. The kernel of a homomorphism is a submodule.
3. If T is a homomorphism from an R – module M to an R – module N , then T is one – one iff
K(T) = {0}.
4. If M is an R – module and N is a R – submodule of M. Consider the map T : M M N defined
by T(m) = m N for all m M . Then prove that T is a homomorphism of M onto M N and K(T)
= N where K(T) is kernel of T.
5. The submodule of the quotient module M N are of the form S N where S is a submodule of M
containing N.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 4
6. Fundamental theorem of homomorphism on modules : Let T is a homomorphism of an R – module
M onto an R – module N with K(T) = A. Then M A is isomorphic to N i.e. M K(T) N .
Def. 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 S and we write ;
A = <S>
It is clear by definition that
(i) < S > is the smallest submodule of M containing S.
(ii) < S > is the intersection of all submodules of M containing S.
7. Show that the submodule of a unital R – module M generated by a subset S of M consists of all
linear combinations of S i.e. L(S) = < S >.
Proof : Let L(S) denote the set of all linear combinations of the elements of S i.e.
L(S) = {r1a1 r2 a2 .......... rn an : ri R , ai S}
First we shall prove that L(S) is a submodule of M
Let a = r1a1 r2 a2 .......... rnan
b = s1b1 s2b2 .......... smbm be any two elements of L(S)
where ri , si R and ai , bi S for all i,
We have
a b (r1a1 ........... rn an ) ( s1 )b1 ( s2 )b2 .......... ( sm )bm
is a linear combination of elements of S
a b L(S)
L(S) is an additive subgroup of M
If r R and a r1a1 .......... rnan L(S) then
ra r (r1a1 r2 a2 .......... rn an )
= (rr1 )a1 (rr2 )a2 ........... (rrn )an L(S)
Hence L(S) is a submodule of M. Now , let a S , then a S , 1 R
a.1 L(S) a L(S)
Hence S L(S) i.e. 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 , since W is
closed under addition and scalar multiplication.
L(S) W
Hence , L(S) is smallest submodule of M containing S
L(S) = < S >.
Def. Finitely generated module : An R – module M is said to be finitely generated if it is generated
by some finite subset of M. Thus M is finitely generated if there exists elements
a1 , a2 ,.........., an M such that each element m M can be expressed :
m r1a1 r2 a2 ......... rnan where r1 , r2 ,......., rn R
In this case we write :
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 5
M = a1 , a2 ,.........., an
Def. Cyclic module : An R – module M is said to be cyclic if there exists an element m0 M s.t.
M = m0 {rm0 : r R}
Thus a cyclic submodule is generated by a single element.
Def. Irreducible Sub – module : An R – module M is said to be irreducible if its only submodules are
{0} and M. In other words , if it has no proper submodule.
Theorem : Prove that any unital , irreducible R – module is cyclic .
Proof : Let M be unital and irreducible R – module , so the only submodules of M are {0} and M .
We claim that M is cyclic. If M = {0} , then obviously M is cyclic. Let M {0} then there exist
at least one element m0 M s.t. m0 0.
Let A = {rm0 : m0 M , r R} we shall prove that A is a submodule of M.
Let a , b A , then
a r1m0 and b r2 m0 for r1 , r2 R and m0 M
then , a b r1m0 r2 m0
= (r1 r2 )m0 A
A is an additive subgroup of M.
Now let r1 R and a rm0 A then r1a r1 (rm0 ) (r1r )m0 A
A is a submodule of M.
Now , since M is unital , so
1.m0 m0 where 1 is unity of R .
But 1 R , m0 M 1.m0 A m0 A . Since m0 0 A {0}
But M is irreducible , so we must have A = M
But A is cyclic M is cyclic.
Theorem : Let A and B be R-sub modules of R-module M and N respectively.
MN M N
Then
AB A B
M N
Proof : Define a mapping f : M N by setting
A B
f (m , n) = (m + A , n + B) for all m M , n N.
We claim that f is an onto homomorphism.
(i) f is well-defined : Let (m , n) and (x ,y) be any two elements of M N such that (m , n) = (x ,
y)
m = x and n = y m + A = x + A and n + B = y + B
(m + A , n + B) = (x + A , y + B) f (m , n) = f (x , y)
f is well defined.
(ii) f is R-module homomorphism : Let (m , n) , (x , y) M N and r R ,
Then f ((m , n) + (x , y)) = f (m + x , n + y)
= (m + x + A , n + y + B)
= ((m + A) + (x + A) , (n + B) + (y + B))
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 6
= (m + A , n + B) + (x + A , y + B)
= f ((m , n)) + f ((x , y))
Also, f (r(m , n)) = f (rm , rn)
= (rm + A , rn + B)
= r(m + A , n + B)
= r f ((m , n))
Hence, f is a R-module homomorphism.
M N
(iii) f is onto : Let X , then X = (m + A , n + B) for some m M , n N.
A B
Thus for (m , n) M N, we have f (m , n) = (m + A , n + B) = X f is onto.
M N
Thus by fundamental theorem of module homomorphism, M N Kerf
A B
......(1)
Now Kerf = {(m , n) : (m + A , n + B) = ( A , B)} = {(m , n) : m A , n B} = A B
MN M N
Using this in (1) , we get
AB A B
Example : Let M be an R-module and let End R (M) be the set of all R-homomorphisms of M into
M. Make appropriate definitions of addition and multiplication of elements of so that End R (M)
become a ring.
Solution : We define addition and multiplication as follows
(a) For S , T End R (M) , define (S + T) (m) = S(m) + T(m)
(b) For S , T End R (m) , define (SoT) (m) = S (T(m))
Now we claim that End R (M) is a ring under these operations.
(i) Addition is closed : Let S , T End R (M) . We claim that (S + T ) is a R-module homomorphism.
We have for m1 , m2 N , (S + T) (m1 + m2) S m1 m2 T m1 m2
= S(m1) + S(m2) + T (m1) + T(m2)
= S(m1) + T(m1) + S(m2) + T(m2)
= (S + T) (m1) + (S + T) (m2)
Also for r R , m M , we have ,
(S + T) (rm) = S (rm) + T (rm)
= r S(m) + r T (m)
= r (S (m) + T (m))
= r (S + T ) (m)
Hence S + T is a R-module homomorphism. Therefore S + T End R (M)
(ii) Associativity : Let T , S, H End R (M) then for m M , we have
(T + (S + H)) (m) = T(m) + (S +H) (m)
= T(m) + (S(m) + H(m))
= (T(m) + S(m)) + H(m)
= (T + S) (m) + H(m)
= ((T + S) + H ) (m) [ T(m) , S(m) , H(m)
M]
So, + is associative in End R (M) .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 7
(iii) Existence of identity : Define O : M M by setting O (m) = 0 for all m M.
Then O is R-homomorphism.
For, O (m1 + m2) = 0 = 0 + 0 = O (m1) + O (m2)
and O (rm) = 0 = r.0 = r O (m)
O End R (M) .
Let T End R (M) , then (T + O )(m) = T(m) + O (m) = T(m) + 0 = T(m) T + O = T
Also, O + T = T O is additive identity in End R (M) .
(iv) Existence of inverse : Let T End R (M) T : M M is R-module homomorphism.
Define H : M M by setting H(m) = T(m) m M.
Then, H(m1 + m2) = T(m1 + m2)
= T(m1) T(m2)
= H(m1) + H(m2)
And H(rm) = T(rm) = rT(m) = r( T(m)) = r H(m) H End R (M)
Also, (H + T) (m) = H(m) + T(m) = T(m) + T(m) = 0 H+T = O
Also, T + H = O . Hence H is additive inverse of T.
(v) Addition is commutative : Let T , S End R (M) .
Then (T + S)(m) = T(m) + S(m)
= S(m) + T(m) = (S + T) (m) [ M is abelian]
T+S = S+T End R (M) is abelian.
(vi) ‘o’ is closed : Let S , T End R (M) . We claim that SoT End R (M)
Let m1 , m2 M. Then (SoT) (m1 + m2) = S(T(m1 + m2))
= S(T(m1) + T(m2))
= S(T(m1)) + S(T(m2))
= (SoT)(m1) + (SoT) (m2)
Also, (SoT)(rm) = S(T(rm)) = S(r T(m) = r(S(T(m)) = r(SoT(m))
SoT End R (M)
(vii) ‘o’ is associative : Since composition of mappings is always associative, so ‘o’ is associative.
(viii) Distributivity : Let T , S , H End R (M) .
We claim that To(S + H) = ToS + ToH and (S + H)oT = SoT + HoT.
Let m N , then (To(S + H))(m) = T ((S + H) (m))
= T(S(m) + H(m)
= (ToS) (m) + (To H)(m)
Similarly, (S + H)oT = SoT + HoT . So, End R (M) is a ring under two operations.
Remark : In the proof of Schur’s lemma before giving the main proof , we may first prove that
End R (M) is actually a ring as proved above.
Schur’s Lemma : If M is an irreducible (simple) R-module, then the ring of endomorphisms
End R (M) is division ring (or skew-field).
Proof : We have already proved that End R (M) is a ring. To show that End R (M) is a division ring
, we shall prove that every non-zero element of End R (M) is invertible.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 8
Since M is given to be irredueible so it has only two sub-modules, M itself and {0}.
Let ( 0) End R (M) . We consider the R-sub modules Ker and (M). If Ker = M , then
= 0, a contradiction.
Thus Ker = {0} , so is one-one. Further, if (M) = {0} , then = 0 , a contradiction.
Thus (M) = M , so is onto. Hence is one-one and onto, which proves that is invertible.
Remark : Sometimes the endomorphism ring End R (M) is also denoted by Hom R (M , M) .
Definition : Elements x1 , x2 ,…….., xn of an R-module N are called linearly independent if, for
any a1 , a2 ,…….., an R,
a1 x1 +a2x2 +……..+anxn = 0 implies a1 = a2 =…….= an = 0.
These elements are called linearly dependent if they are not linearly independent.
Definition : An R-module N is said to be generated by a set { x1 , x2 ,…….., xn} if every element of
N can be expressed as a linear combination of x1 , x2 ,…….., xn i.e. if x N, then
x = a1 x1 +a2x2 +……..+anxn , ai R.
Definition : A sub set B of an R-module N is said to be a basis if
(i) N is generated by B.
(ii) B is linearly independent set.
Free Module : An R-module N is said to be a free module if N has a basis. In other words N is
free module if there exist a subset B of N such that N is generated by B and B is linearly
independent.
Theorem : Let N be a free R-module with a basis {e1 , e2 ,…….., en}. Then N Rn .
Proof : Since { e1 , e2 ,…….., en} is a basis of N, every element of N is a linear combination of
ei s . Let { f1 , f2 ,…….., fn} denotes the standard basis of Rn , where fi = (0 ,…….., 1 , 0 ,……..,0)
where 1 is on the ith position. Then every element of Rn is linear combination of fi’s. We define a
mapping : N Rn by setting
(r1e1 + r2e2 +……..+ rnen) = r1 f1+ r2 f2 +……..+ rn fn.
n n
(i) is well defined : Let re
i 1
i i r 'e
i 1
i i
(r1 r1' )e1 + (r2 r2' )e2 + ……..+ (rn rn' )en = 0
r1 r1' = r2 r2' = …….. = rn rn' = 0 [ eiS are L. I.]
r1 = r1' , r2 = r2' , …….. , rn = rn'
n n
ri fi
i 1
r' f
i 1
i i
n n
re
i i ri'ei
i 1 i 1
is well defined.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 9
n n
(ii) is homomorphism : Let m = rei i
i 1
and m' r 'e
i 1
i i
n n
n n
Then (m + m ) = ri ri' ei
i 1
i 1
ri ri' f i = ri fi
i 1
r' f
i 1
i i (m) (m)
n n n
and (rm) = rre
i 1
i i
rri f i = r
i 1
r f
i 1
i i r (m)
is homomorphism.
n n
(iii) is one-one : Let m = rei i and m
i 1
r 'e
i 1
i i be two elements of N such that.
(m) = ( m )
n n
ri fi
i 1
r' f
i 1
i i
n
(r r ' ) f
i 1
i i i 0
ri ri' 0 , 1 i n [ fi ’s are L.I.]
ri ri' , 1 i n
n n
ri ei
i 1
r e
i 1
i i
m m .
n
(iv) is onto: Let r i fi be any element of R n where ri R , 1 i n.
i 1
n
Then r e
i 1
i i is clearly pre-image of the chosen element. Hence is an isomorphism and so N
Rn .
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 s.t. xi is a linear
combination of elements of Bi with coefficients in R.
n
Let S = B ,
i 1
i then clearly S is finite since each Bi is finite.
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. Therefore, B is finite since S is finite.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 10
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 Rn Rm Rn .
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
i.e., we have
(e1) = a11 f1 + a21 f2 +……..+an1 fn
(e2) = a12 f1 + a22 f2 +……..+an2 fn
…………………………………….
(em) = a1m f1 + a2m f2 +……..+anm fn
a11 a12 a1m
a a22 a2 m
Then, matrix of = A = 21
an1 an 2 anm nm.
Again, (fi) R m , 1 j n, so let us write
(fi) = b1j e1 + b2j e2 +……..+bmj em
i.e 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.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 11
b11 b12 b1n a11 a12 a1m 1 0 0
b b22 b2 n a
a22 a2 m 0 1 0
i.e. 21 21 =
bm1 bm 2 bmn m n . an1 an 2 anm n m. 0 0 1 m m
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 bkj =
ik ……(2)
k 1 0 if i j
B
Let A = [A O] and B = be n n augmented matrices, where each of the O blocks
O
is a matrix of appropriate size i.e.
b11 b12 b1n
a12 a1m 0 0 b b22 b2 n
a11 21
a a22 a2 m 0 0
A = 21 , B =
bm1 bm 2 bmn
0
an1 an 2 anm 0 0 nn 0 0
0 0 0 nn.
Form these two matrices we note that
1 0 0
0 1 0
A B = = In [Using (2)]
0 0 1 nn
1 0 0 0 0
0 1 0 0 0
Im 0
And BA = [Using (1)]
0 01 0 0 0 0
0 0 0 0 0
0 0 0 0 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 12
So, we have
I 0
det (AB) = det (In) = 1 and det (BA) = det m = 0
0 0
det (AB) det (BA) …(3)
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 i.e. all basis of N have the same number of elements.
Definition : 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.
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 re
i i ri xi
i 1 i 1
n n
(i) is well-defined : Let x = rei i and y = r 'e i i be two elements of R n s.t. x=y
i 1 i 1
n
i.e. (r r ' ) e
i 1
i i i 0 ri ri' 0 , 1in [Since ei' are L.
I.]
n n
ri ri' , 1in ri xi
i 1
r 'x
i 1
i i ( x) ( y) .
(ii) is homomorphism :
n n
Let x = rei i
i 1
, y = r 'e
i 1
i i and r R, then
n n
(x + y) = (ri ri' ) ei (r i ri' ) xi
i 1 i 1
n n
= r x r 'x
i 1
i i
i 1
i i = (x) + (y)
n
n
and (rz) = rre
i 1
i i
= rr x
i 1
i i r ( x) .
n
(iii) is onto : Let m = rx
i 1
i i be any arbitrary element. Then ri R and consider the element
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 13
n n
n
x= re i i R n . Then (x) = re i i = ri xi = m. So, x is pre-image of m. is
i 1 i 1 i 1
onto. Hence, ( R n ) = N i.e. N is homomorphic image of R n which is a finitely generated free-
module.
Theorem : Every finitely generated module is isomorphic to a quotient group of a finitely generated
free module.
Proof : Reproduce proof of above theorem and then by fundamental theorem of module
homomorphism
R n ker M.
i.e. N is isomorphic to a quotient group of finitely generated free-module.
Results : (1) Every finitely generated module is isomorphic to a quotient group of a finitely
generated free module i.e. if N is a finitely generated R-module, then N R n K where K is a
sub-module of R n .
(2) Let R be a Principal Ideal Domain (P I D) and let E be a free R-module with a basis
consisting of n elements (i.e. F R n ) , then any sub module K of F is also free with a basis
consisting of m elements such that m < n (i.e. K R m m n).
(3) If A is an m n matrix over a P I D R, then A is equivalent to a matrix that has the
diagonal form.
a1
a2 O
ar where ai 0 and a1 a2 ........... ar
0
O
0
In fact, there exist invertible matrices P and Q of order m m and n n , respectively such that
PAQ = diag (a1 , a2 , …….., ar ,0 , 0……..0) where a1 a2 ............ ar
Fundamental Structure Theorem (or Decomposition Theorem) of finitely generated modules
over Principal Ideal Domain :
Let R be a P I D 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 .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 14
Now, since R n is a free R-module where R is a P I D and K is a sub module of R n , so we
must have
K R m where m n [By Result (2)]
m m
Let be this isomorphism from R to K i.e., K = ( R ). Let { e1 , e2 , …….., em} be a basis
of R m . Let us write
a11
a
(e1 ) R n
21
an1
a1m
a2 m
(em ) R n
anm
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 resp. s.t.
P A Q = 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
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 15
M R Ra1 R Ra2 ...... R Rar R s
Application to finitely generated abelian groups : Since the ring of integers is a P I D 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 .
Smith-Normal form of a matrix : If A is m n matrix over a PID R, then A is equivalent to a
matrix that has the diagonal form.
a1
a2 O
ar
0
O
0
where ai 0 and a1 a2 ............ ar . This diagonal form equivalent to A is known as Smith-normal
form of A and number of non-zero elements gives the rank of A.
Invariant Factors : The non-zero diagonal elements in the Smith-normal form of a matrix A are
called the invariant factors of A.
Elementary operations : In obtaining the Smith normal form of a matrix A over a PID R we shall
use the following operations known as elementary row (column) operations
(i) Interchanging the two rows (columns) we denote the operation of interchanging the ith and jth
row (column) by Ri Rj (Ci Cj).
(ii) Multiplying the elements of one row (column) by a non-zero element of R. The operation of
multiplying the ith row (column) of A by is denoted by Ri ( Ci).
(iii) Adding to the elements of one row (column) times the corresponding elements of a
different row (column) The operation of adding to the elements of ith row (column) times the
corresponding elements of the jth row (column) is denoted by Ri + Rj (Ci + Cj).
Example 1: Obtain the Smith-normal form and rank for the matrix with integral entries
1 2 3
4 .
5 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 16
1 2 3
Solution : Let A = 4
5 0
1 2 3
Operating R2 4R1 , A ~ 0 3 12
1 0 0
Operating C2 2C1 , C3 3C1, ~ 0 3 12
1 0 0
Operating C3 4C2 , ~ 0
3 0
1 0 0
Operating ( 1) R2, ~ 0
3 0
which is the required Smith-normal form. Clearly rank (A) = 2
Example 2 : Obtain the Smith-normal form and rank for the matrix A over Z, where
0 2 1
A = 3 8 3
2 4 1
Solution : we have
0 2 1
A = 3 8 3
2 4 1
0 1
2
Operating R3 R1, A ~ 3 8 3
2 6 0
0 1
2
Operating R2 + 3 R1, ~ 3 14 0
2 6 0
0 2 1
Operating C2 + 3 C1 A ~ 3 5 0
2 0 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 17
0 0 1
Operating C2 + 2C3 A ~ 3 5 0
2 0 0
0 0 1
Operating R2 2 R3 A ~ 1 5 0
2 0 0
0 0 1
Operating R3 + 2R2 A ~ 1 5 0
0 10 0
0 0 1
Operating C2 5C1 A ~ 1 0 0
0 10 0
1 0 0
Operating R1 R2 A ~ 0 0 1
0 10 0
1 0 0
Operating C2 C3 A ~ 0 1 0
0 0 10
1 0 0
Operating ( 1)R2 , ( 1) R3 A ~ 0 1 0
0 0 10
which is required Smith-normal form of A, Rank (A) = 3.
Example : (3) Obtain Smith-normal form of matrix A over Q[x] , where
x 3 2 0
A = 1 x 1 .
1 3 x 2
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 18
x 3 2 0
Solution : A = 1 x 1
1 3 x 2
1 x 1
Operating R1 R2 A ~ x 3 2 0
1 3 x 2
1 x 1
Operating R3 R1 ~ x 3 2 0
0 3 x x 3
1 x 0
Operating C3 C1 ~ x 3 2 x3
0 3 x x 3
1 x 0
Operating R2+R3 ~ x 3 x 1 0
0 x 3 x 3
1 x 0
2
Operating R2+(x+3) R1 , ~ 0 ( x 1) 0
0 x3 x 3
1 0 0
2
Operating C2 + x C1 ~ 0 ( x 1) 0
0 x3 x 3
1 0 0
2
Operating C2 + C3 ~ 0 ( x 1) 0
0 6 x 3
1 0 0
Operating R2 R3 ~ 0 6 x 3
0 ( x 1)2 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 19
1 0 0
1
Operating R2 ~ 0 1 ( x 3) 6
6 0 ( x 1)2 0
1 0 0
x 3
Operating C3 C2 ~ 0 1 0
6 0 ( x 1) 2 ( x 3) ( x 1) 6
2
1 0 0
2
Operating R3 + ( x 1) R2 ~ 0 1 0
0 0 ( x 3) ( x 1) 6
2
1 0 0
Operating 6(R3) ~ 0 1 0
0 (x +1) 2 2
(x +3) (x +1)
which is required Smith-normal form.
Example 4 : Obtain the Smith-normal form of A over Q[x] where
x 4 2
A = 3 8 x 3
4 8 2 x
Solution : we have
x 4 2
A = 3 8 x 3
4 8 2 x
4 8 2 x
Operating R1 R3 A ~ 3 8 x 3
x 4 2
4 8 2 x
x
Operating R3 + R1 ~ 3 8 x 3
4 2
0 x x
4 2x 2
2 4
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 20
4 8 2 x
3 0 3 3x
Operating R2 + R1 ~ 2 x
4 2 4
0 x x2
4 2x 2
2 4
4 8 2 x
0 3 3x
Operating R3 2R2 ~ 2 x
2 4
0 x2
0 5 x
4
4 0 2 x
0 3 3x
Operating C2 + 2C1 ~ 2 x
2 4
0 x2
0 5 x
4
1 0 2 x
1 0 3 3x
Operating (C1) ~ 2 x
4 2 4
0 x2
0 5 x
4
1 0 2 x
3
Operating C3 (C2) ~ 0 2 x 0
4 2
0 x
0 5 x
4
1 0 0
Operating C3 + (2 + x)C1 ~ 0 2 x 0
x2
0 0 5 x
4
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 21
1 0 0
Operating ( 4) R3 ~ 0 2x 0
0 0 x 2 4 x 20
1 0 0
Operating C2 C3 , R2 R3 ~ 0 x 2 4 x 20 0
0 0 2 x
1 0 0
2
Operating R2 + R3 ~ 0 x 4 x 20 2 x
0 0 2 x
1 0 0
Operating C2 + xC3 ~ 0 2 x 20 2 x
0 2 x x2 2 x
1 0 0
Operating C2 2C3 ~ 0 16 2 x
0 x2 4 x 4 2 x
1 0 0
1
Operating (R 2 ) ~ 0 1 2 x 16
16
0 x2 4 x 4 2 x
1 0 0
2 x
Operating C3 C2 ~ 0 1 0
16 2
0 [16 x 4 x 4]
x2 4 x 4 (2 x)
16
1 0 0
Operating R3 ( x2 +4x 4)R2 ~ 0 1 0
2
0 ( x 4 x 20)
0 (2 x)
16
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 22
1 0 0
Operating ( 16)R3 ~ 0 1 0
0 0 ( x 2) ( x 2 4 x 20)
which is the required Smith-normal form of A. The invariant factors of A are 1 , 1, (x 2)
x 2 4 x 20 .
Example 5 : Find the invariant factors of matrix A over Q[x], where
5 x 1 2 4
0 5 x 2 2
A =
0 0 5 x 3
0 0 0 4
Solution : we have
5 x 1 2 4
0 5 x 2 2
A =
0 0 5 x 3
0 0 0 4
1 5 x 2 4
5 x 0 2 2
C1 C2 A ~
0 0 5 x 3
0 0 0 4
C2 (5 x) C1 , C3 +2C1, C4 4C1
1 0 0 0
5 x (5 x )2 12 2 x 4 x 18
A ~
0 0 5 x 3
0 0 0 4
1 0 0 0
2
0 (5 x) 12 2 x 4 x 18
R2 (5 x)R1, A ~
0 0 5 x 3
0 0 0 4
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 23
1 0 0 0
0 4 0 0
R2 R4 , C2 C4 A ~
0 3 5 x 0
0 18 4 x 12 2 x (5 x) 2
1 0 0 0
0 1 0 0
1
(R2) A ~
4 0 3 5 x 0
0 4 x 18 12 2 x (5 x )2
R3 3R2 , R4 (4x 18)R2
1 0 0 0
0 1 0 0
A ~
0 0 5 x 0
0 0 12 2 x (5 x) 2
1 0 0 0
0 1 0 0
1
R3 R4 A ~ (5 x )2
2 0 0 1
2
2
0 0 12 2 x (5 x)
1 0 0 0
0 1 0 0
( 1)R3 A ~ (5 x) 2
0 0 1
2
0 0 12 2 x (5 x ) 2
1 0 0 0
0 1 0 0
(5 x )2
C4 + C3 A ~ 0 0 1 0
2
0 2 12 2 x
0 12 2 x (5 x) 1
2
1 0 0 0
0 1 0 0
R4 (12 2x) R3 A ~
0 0 1 0
0 0 0 (5 x )2 (6 x 1)
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 24
1 0 0 0
0 1 0 0
=
0 0 1 0
3
0 0 0 (5 x)
Thus, the invariant factors of A are 1 , 1 , 1 , (5 x )3 .
Example 6 : Find the rank of the sub-group of Z4 generated by following elements
(i) (3 , 6 , 9 , 0) , ( 4 , 8 , 12 , 0)
(ii) (2 , 3 , 1 , 4) , (1 , 2 , 3 , 0) , (1 , 1 , 1 , 4)
(iii) ( 1 , 2 , 0 , 0) , (2 , 3 , 1 , 0) , (1 , 1 , 1 , 1)
Solution : (i) It is clear that 4(3 , 6 , 9 , 0) = 3( 4 , 8 , 12 , 0) i.e. the two elements are
linearly dependent over Z and hence rank = 1.
2 3 1 4
(ii) The corresponding matrix is A = 1 2 3 0
1 1 1 4
1 2 3 0
Operating R1 R2 A ~ 2 3 1 4
1 1 1 4
1 2 3 0
Operating R2 2R1 , R3 R1 ~ 0 1 5 4
0 1 2 4
1 0 0 0
Operating C2 2C1 , C3 3C1 ~ 0 1 5 4
0 1 2 4
1 0 0 0
Operating ( 1) R2 ~ 0 1 5 4
0 1 2 4
1 0 0 0
Operating R3 + R2 ~ 0 1 5 4
0 0 3 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 25
1 0 0 0
Operating C3 5C2 , C4+4C2 ~ 0 1 0 0
0 0 3 0
Clearly rank = 3
1 2 0 0
(iii) The corresponding matrix is A = 2 3 1 0
1 1 1 1
1 1 1 1
Operating R1 R3 A ~ 2 3 1 0
1 2 0 0
1 1 1 1
Operating R2 2R1 , R3 + R1 ~ 0 5 1 2
0 3 1 1
1 0 0 0
Operating C2 C1, C3 C1,C4 C1 ~ 0 5 1 2
0 3 1 1
1 0 0 0
Operating C2 C3 , R2 R3 ~ 0 1 3 1
0 1 5 2
1 0 0 0
Operating R3 + R2 ~ 0 1 3 1
0 0 2 1
1 0 0 0
Operating C3 3C2, C4 C2 ~ 0 1 0 0
0 0 2 1
1 0 0 0
1 0
Operating C4 C3 ~ 1 0 0
2
0 0 2 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-I PAGE 26
1 0 0 0
Operating ( 1)R3 ~ 0 1 0 0
0 0 2 0
Rank = 3
Rational Canonical Form
Rational canonical block of invariant factor
0 0 0 a0
a
1
a0 a1 x a2 x 2 .......... ak 1x k 1 x k is
0 0
0 0 1 ak 1 kk
Working rules to find rational canonical form of a matrix of order n n .
(i) Find the invariant factor A xl.
(ii) Construct the rational block of each invariant factor.
(iii) Arrange these blocks in the increasing order of their size on the diagonal of n n matrix to get
the rational canonical form.
3 2 0
Example 7 : Let A = 1 0 1 . Find the rational canonical form of this matrix over Q.
1 3 2
3 x 2 0
Solution : Here A xl = 1 x 1
1 3 2 x
Let us find the Smith normal form of A–xl.
1 0 0
We get 0 1 0
0 0 ( x 3)( x 1)2
So, the invariant factor of A are 1 , 1 and (x + 3) (x + 1)2 = 3 + 7x + 5x2 + x3
0 0 3
Therefore, rational canonical form of A is 1 0 7 .
0 1 5
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 1
Def. Noetherian Module : An R-module M is said to be Noetherian if for every ascending sequence
of R – submodules of M
M1 M 2 M 3 ......
there exists a positive integer k such that M k = M k + 1 = M k + 2 = ...........
If M is Noetherian, then we also say that M has ascending chain condition (acc).
Def. Artinian Module : An R-module M is said to be Artinian if for every descending sequence of
R – submodules of M
M1 M 2 M 3 .............
there exists a positive integer k such that M k = M k + 1 = M k + 2 = .............
If M is Artinian, then we also say that M has descending chain condition (dcc).
Example 1 : Show that as a – module is Noetherian but not Artinian.
Solution : Since the ring of integers is a principal ideal domain, any ascending chain of ideals of
is of the form
< n > < n1 > < n2 > ..........., n, n1 , n2
But, < ni > ni 1 iff ni 1 / ni , so any ascending chain of ideals (or sub modules of ) starting
with n can have at most n distinct terms. Hence every properly ascending chain of sub modules of
terminates after a finite number of steps. So is Noetherian as a -module.
But as a -module has an infinite property descending chain
< n > < n2 > < n3 > ..............
showing that is not Artinian as a -module.
Example 2 : Let V be an n – dimensional vector space over a field F. Then V is both Noetherian
and Artinian. Because , if W is a proper subspace of V , then dim W < dim V = n, which implies
that any properly ascending or descending chain of subspaces can not have more than n +1 terms.
m m
Example 3 : Let p be prime number , and let R = ( p ) n : 0 n 1 be the ring
p p
where addition is module positive integers and multiplication is defined as ab 0 for all a , b R,
then
1 2 p k 1
(i) Each ideal in R is of the form A k 0 , k , k ,....., where k is some positive integer.
p p pk
(ii) R is Artinian but not Noetherian.
Solution : (i) Let A {0} be any ideal of R, and let k be the smallest positive integer such that
m n
for some positive integer m, k A. Consider i with i k and (n , p) = 1
p p
n n
We claim that i A. Let , if possible, i A
p p
n n n
i
i ........ i A ( p i--k times)
p p p
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 2
n. p i--k n
i
k A
p p
......(1)
1
Also by choice of k, A
p k --1
......(2)
Now, because gcd (n , p) = 1 , there exist integers a and b such that an + bp = 1
n n n n na
Now, by (1), k
A implies that k
k ........ k (a times) A A
p p p p pk
......(3)
p
Also, by (2), A
pk
p p p bp
k
k ......... k A (b times) A ……(4)
p p p pk
na bp 1 1
By (3) and (4) , we have, k
k
A which is a contradiction, because if A
p p pk
then
1 1 1 m
k
k ........ k k A , which is not so.
p p p p
n 1 2 p k --1 1
Thus, A for i k , (n , p) = 1 and so A 0 , , ,........,
pi p k --1 p k --1 p k --1
This ideal is denoted by A k 1 .
(ii) Because each ideal contains a finite number of elements, each descending chain of ideals must be
finite. Hence every properly descending chain of sub modules of R terminates after a finite number
of steps. Hence R is Artinian as a R-module.
On the other hand, we see that the chain
A1 A2 A3 .............
is an infinite properly ascending chain of ideals. So R is not a Noetherian module as R-module.
Hence R as a R-module is Artinian module but not Noetherian module.
Theorem 1 : 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 .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 3
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 given to be Noetherian.
Hence the family must have a maximal element.
(ii) (iii)
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 sub – modules of M.
Let us denote Nk = < a1 , a2 ,……., ak > , then N1 N2 ……. Nk ........
Let family of all these sub-modules be i.e. = {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 sub-module 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.
Let N = N i . We claim that N is also a sub-module of M.
i
Let x , y N
i
i and r R. Then x Nr and y Ns for some integers r and s.
Because either Nr Ns or Ns Nr, both x and y lie in one sub-module Nr or Ns , and
hence x y and r x lie in same sub-module. But both Nr and Ns are contained in N, so x y
N and r x N, and hence N is a sub-module 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 i.e., a1 , a2 ,……., an
Nk
But N = < a1 , a2 ,……, an > , so N is smallest sub-module containing a1 , a2 ,……., an .
This implies N Nk. . Also Nk N. Hence N = Nk
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 4
But then Nk = Nk +1 = Nk +2 = .........., therefore, M is a Noetherian R-module.
Def. Finitely co-generated module : An R-module M is said to be finitely co-generated if, for
each family M of sub-modules of M for which M = {0} we
must have
M
'
= {0} , for some finite subset of .
Theorem 2 : 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 M 3 .
If this process continues indefinitely (i.e. 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.
(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 i.e., 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
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 5
A M
A M
= N [By
(1)]
n n
Again, N = M
M i = A . Hence A = N =
i =1
M
i =1
i
n
n
Now, M i N = M i N = N N = {N}.
i =1 i =1
n 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)
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 = Mi . Then N is a sub-module of M and N Mi for all i .
i
Now, consider the family M i Ni of sub-modules of M N . We see that
i i
M
N = M i 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 ni N
i =1
of M i Ni such that M
i =1
ni N = {N}
Let k = max. {n1 , n2,........,nr} then
r
M
i 1
ni Mk [Since the chain is descending]
r
r
Now {N} =
i 1
ni M
N M n 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.
Theorem 3 : Let M is a Noetherian R-module. Then every sub-module and quotient module of M
are also Noetherian.
Proof : 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 N, be any sub-
module of N then it is also finitely generated i.e every sub-module of N 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
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 6
sub-module of M N is finitely generated. So let A 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
= r1 (x1+N) + r2 (x2 +N) +..........+ rn (xn + N)
Hence A N = < x1+N, x2 +N, ..........,xn+N > i.e. A N is finitely generated.
Therefore, M N is Noetherian.
Theorem 4 : If N is a sub-module of M such that both N and M N are Noetherian then so is
M.
Proof : 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 .
But M N is Noetherian , so A + N N must be finitely generated.
Now , we know that A + N N A AN
A A N is also finitely generated.
Also A N, being sub-module of a 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
x + A N = r y
j 1
j j A N , rj R
n
= r y
j 1
j j AN
n
x r y j j AN AN
j 1
n
x rj y j A N
j 1
But A N = < x1 , x2 ,.......,xm >
n
So, x rj y j = s1x1 + s2x2 +........+ smxn , si R
j 1
n m
x r y
j 1
j j s x
i 1
i i
A = < x1 , x2 ,.....,xm , y1 , y2 ,.......,yn >
Hence A is finitely generated and therefore M is Noetherian R-module.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 7
Theorem 5 : Let M is a Artinian R-module. Then every sub-module and quotient module of M
are also Artinian.
Proof : (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 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 .
i.e., M1 N M 2 N M 3 N ......... M k N ..........
Here, Mi are sub-modules of M and since M i N M i +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 a positive integer k s.t. Mk = Mk +1 = Mk +2 = .......... and then, we
have
M k N = M k +1 N = M k +2 N = ..........
Hence, M N is a Artinian R-module.
Theorem 6 : If N is a sub-module of M such that both N and M N are Artinian then so is M.
Proof : 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 A 2 N N ........... A n N N ..........of sub-
modules of M N .
But M N is Artinian, so there exist a positive integer r such that
A r N N = A r +1 N N = A r +2 N N = .........
i.e., A r N N = A r +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 Ai1 N for all i. Hence, we have a descending chain
A1 N A2 N ........ An N ......... of sub-modules of N.
But N is artinian, so there exist a positive integer s such that As N = As +1 N = As +2 N =
.........
i.e., As N = As+i N i 0
……(2)
Let k = max{r , s} , then by (1) and (2), we must have
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 8
Ak +N = Ak + i + N i 0 ......(3)
and Ak N = Ak+i N i 0 ......(4)
We claim that Ak = Ak + i 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)]
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 i 0.
Therefore, the considered chain of sub-modules of M becomes stationary after a finite number of
steps.
Hence M is Artinian.
Theorem 7 : Every homomorphic image of Noetherian module is again Noetherian.
Proof : To prove the theorem, we first prove that quotient module of a Noetherian R-module M is
again Noetherian. (Give proof here). Now, let P be any homomorphic image of M, then there
exist one-one and onto homomorphism, say : M P. Then by fundamental theorem of
module homomorphism,we have,
M ker P
But M ker is Noetherian since M is given to be Noetherian and we have proved above that
quotient module of a Noetherian module is again Noetherian. Therefore, P is also Noetherian.
Theorem 8 : Every homomorphic image of a Artinian module is again Artinian.
Proof : First prove that quotient module of a Artinian module is Artinian and then apply fundamental
theorem of homomorphism as in above theorem.
Result : Let M1 and M2 are R-modules and N1 and N2 are its sub-modules, then N1 N2 is a
sub-module of M1 M2 and M1 M 2 N1 N 2 M1 N1 M 2 N 2 and, in general,
if N1 , N2 ,........, Nk are sub-modules of M1 , M2 ,..........., Mk then
M1 M 2 ........ M k N1 N 2 .......... N k M1 N1 M 2 N 2 .......... M k N k
Theorem 9 : If M1 , M2 ,..........., Mk are Noetherian then so is M1 M2 ........... Mk.
Proof : We shall prove the result by induction on k. Let us first suppose that k = 2.
Let M1 and M2 are two Noetherian R-modules, we shall prove that M1 M2 is also Noetherian.
Let us define : M1 {0} M1 by setting
(x , 0) = x (x , 0) M1 {0}
(i) is well – defined : Let (x , 0) , (y , 0) M1 {0} such that (x , 0) = (y , 0)
x = y
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 9
(x , 0) = (y , 0)
(ii) is homomorphism : [(x , 0) + (y , 0)]
= (x + y , 0)
= x + y = (x , 0) + (y , 0)
and [r(x , 0)] = (rx , 0] = rx = r (x , 0)
(iii) is one-one : Let (x , 0) , (y , 0) M1 {0} such that (x , 0) = (y , 0)
x = y
(x , 0) = (y , 0)
(iv) is onto : Let x M1 be any element, then (x,0) M1 {0} and (x , 0) = x
Hence is an isomorphism.
M1 {0} M1
But M1 is Noetherian so M1 {0} is also Noetherian.
M1 M 2 M1 M 2
Now, we see that {M1} M 2 M 2
M1 {0} M1 {0}
But M2 is Noetherian so M1 M 2 M1 {0} is also Noetherian.
Hence M1 M2 is Noetherian , since we know that if N and M N are Noetherian then so is M and
here M1 {0} and M1 M 2 N1 {0} are both Noetherian. Therefore, result is true for k = 2.
Now, suppose that k >2 and result is true for k 1.
Now, M1 , M2 ,........, Mk–1 are Noetherian R-modules, so M1 M2 ......... Mk–1 is also Noetherian.
Now, M1 M2 ......... Mk–1 and Mk are Noetherian, so it follows that M1 M2 ........... Mk is also
Noetherian. Hence, we have proved that direct product of Noetherian modules is again Noetherian.
Theorem 10 : M1 , M2 ,............,Mk are Artinian R-modules then so is M1 M2 ............ Mk.
OR
Direct product of Artinian R-modules is again Artinian.
Proof : Reproduce the above proof by writing Artinian in place of Noetherian.
k
Theorem 11 : Let M1 , M2 ,............, Mk be Noetherian R-modules then M
i 1
i is also Noetherian.
OR
Sum of Noetherian R-modules is again a Noetherian R-module.
Proof : We know that finite direct product of Noetherian R-modules is again Noetherian R-module.
Now each Mi is Noetherian, so M1 M2 .......... Mk is also Noetherian.
We define a mapping : M1 M2 ......... Mk M1+ M2 +............+ Mk by setting
( x1 , x2 ,........,xk) = x1 + x2 +.........+xk , xi Mi.
(i) is homomorphism :
[( x1 , x2 ,.........., xk) + ( y1 , y2 ,......., yk)] = ( x1+y1, x2 + y2 ,............, xk + yk)
= x1+y1+ x2 + y2 +..........+ xk + yk = x1+x2 + ...........+ xk + y1+y2+..........+yk
= ( x1 , x2 ,..........., xk) + ( y1 , y2 ,.........., yk)
and [r( x1 , x2 ,..........., xk)] = ( rx1 , rx2 ,............, rxk)
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 10
rx1 rx2 ....... rxk r x1 x2 ........... xk r x1 , x2 ,.........., xk
is a homomorphism.
(ii) is onto : Let x M1 + M2 +......+Mk be any element. Then x = x1 + x2 +......+ xk , xi Mi
Now, ( x1 , x2 ,............, xk) M1 M2 .......... Mk
and ( x1 , x2 ,.........., xk) = x1 + x2 +........... +xk = x so is onto.
k
Hence M
i 1
i is a homomorphic image of a Noetherian module. Also, we know that homomorphic
image of a Noetherian module is again Noetherian and so M1 + M2 +.............+ Mk is Noetherian.
k
Theorem 12 : Let M1 , M2 ,............, Mk be Artinian R-modules then M
i 1
i is also Artinian.
OR
Sum of Artinian R-modules is again Artinian.
Proof : Reproduce the above proof by writing Artinian in place of Noetherian.
Def. Noetherian ring : A ring R is called a left (or right) Noetherian ring if R regarded as a left
(or right) R-module is Noetherian.
Def. Artinian ring : A ring R is called a left (or right) Artinian ring if R regarded as a left (or
right)
R-module is artinian.
Theorem 13 : Let R be a ring. Then the following conditions are equivalent :
(i) R is Noetherian.
(ii) Every non-empty family of left ideals of R has a maximal element.
(iii) Let A be any left ideal of R. Then A is finitely generated.
Theorem 14 : Let R be a ring. Then the following conditions are equivalent :
(i) R is Artinian.
(ii) Every non-empty family of left ideals of R has a minimal element.
(iii) Let A be any left ideal of R. Then R A is finitely co-generated.
Theorem 15 : Every principal left ideal ring R is a Noetherian ring.
Proof : Let A be any left ideal of R. Since R is principal left ideal ring, so A must be generated
by a single element i.e. A is finitely generated. Hence by above theorem, R is Noetherian.
Example 4 : Give an example to show that a sub ring of a Noetherian ring need not be Noetherian.
Solution : Let us consider the ring, denoted by R, of all 2 2 matrices over the rational numbers .
i.e.
a b
R= : a, b, c, d
c d
Then R is an 4-dimensional vector space over and each left ideal as well as right ideal of R is a
sub-space of R over . Thus any ascending or descending chain of left as well as right ideals can
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 11
not contain more than 5 terms. Thus R is both Noetherian and Artinian ring. Now we consider the
sub ring S of R as
a b
S = = : a , b, c
0 o c
We prove that S is not left Noetherian. We see that for each positive integer k ,
0 m 2k
Ak = : m
0 0
is a left ideal of S and Ak Ak +1. Thus there exist an infinite properly ascending chain of left
ideals of S, namely A1 A2 A3 ............
Hence S is not left Noetherian.
Example 5 : Give an example to show that a sub-ring of a Artinian ring need not be Artinian.
Solution : We consider the ring of rational numbers . Now, is a field and we know that a field
has no proper ideals and so the only descending chain of ideals of is {0}.
So, can not have an infinite properly descending chain of ideals showing that is Artinian. Now
is a sub ring of and is not Artinian. For any positive integer n, the strictly descending
chain
< n > < 2n > < 3n > …………., of ideals of infinite
Def. Product of Ideals : Let A and B be any two left (or right) ideals of a ring R. then we define
AB = aibi : ai A, bi B .
finite
Result : For any two left (or right) ideals A and B of a ring R, the product AB is also an left (or
right) ideal of R.
Def. Power of an ideal : Let A be any left (or right) ideal of a ring R, then we define
A2 = ai bi : ai , bi A
finite
Using induction, we can define An for any positive integer n.
Def. Nilpotent element : An element a R is said to be an nilpotent element if there exists an
positive integer n such that an = 0.
Def. Nilpotent ideal : An left (or right) ideal A of a ring R is said to be left nilpotent (or right
nilpotent) if some positive integer n, we have An = {0}.
Def. Nil ideal : An left (or right) ideal A of a ring R is said to be left nil ideal (or right nil ideal)
if for each a A, there exists positive integer n s.t. an = 0 i.e. each element of A is nilpotent.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 12
Example 6 : Consider M, the ring of all 2 2 upper triangular matrices over integers. Consider the
0 1 2
ideal I of M generated by . Clearly, I is nilpotent ideal since I = [0].
0 0
2
0 1 0 0
This ideal I is also a nil ideal since 0 = and this will make all element of I
0 0 0
nilpotent elements.
Theorem 16 : Prove that in an Artinian ring every nil left ideal is nilpotent.
Proof : Let J be an nil left ideal in an Artinian ring R. We shall prove that J is nilpotent. Let, if
possible, J is not nilpotent i.e. Jk {0} for every positive integer k.
Consider the family { J , J2 , J3 ,.........} of left ideals of R. We know that if R is Artinian then
ervery non-empty family of left ideals of R has a minimal element.
So this family has a minimal element, say B = Jm for some positive integer m.
Then B2 = J2m Jm = B
But B is minimal element and B2 B , so we must have B2 = B.
Consider another family ,
= {A : A is a left ideal of R contained in B such that BA {0}}
Now, B is a left ideal of R contained in B and B.B = B2 = B = Jm {0}[ Jk {0} for
every k]
B
This family also must have a minimal element, so let A be a minimal element of . Then A is an
left ideal of R contained in B such that BA {0}. This implies that there exist an non-zero
element a A
such that B a {0}
........(1)
Now , since A is an left ideal of R so r, a A r R
In particular, b, a A b B i.e., Ba A
……(2)
and B (Ba) = B2a = Ba {0} [By
(1)]
Also, clearly Ba is a left ideal of R. Thus, we have, Ba is a left ideal of R contained in B such
that
B(Ba) {0} Ba
Also, by (2), Ba A. But A is minimal element of , so we must have Ba = A.
Now, a A = Ba implies that a = ba for some b B.
Then b2a = b(ba) = ba = a and b3a = (b2a) = ba = a .
Hence we get
bia = a for all positive integer i
……(3)
Now, b B = Jm J and J is a nil ideal. So b must be a nilpotent element i.e. there exist an
positive integer r such that br = 0. Taking i = r in equation (3), we get
bra = a 0.a = a a = 0
which is a contradiction, since a was a non-zero element. Hence J must be nilpotent ideal of R.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 13
Def. Annihilators : Let S be a non – empty set of ring R. We define
r(S) = {x R : Sx = 0} and l(S) = {x R : x S = 0}
Then r(S) and l(S) are called right and left annihilators respectively of S and they are right and left
ideals respectively.
Result : Ideal generated by an element ‘a’ of a ring R is given as
< a > = {ra + as + xay+ na : r, s, x, y R, n }
(i) If R has unity, < a > = {xay : x , y R}
(ii) If R is commutative, < a > = {na + ra : n , r R}
(iii) If R is commutative with unity, < a > = {ar : r R} = aR = Ra.
Theorem 17 : 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 i.e. 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
Let J = R a R , then, J2 = RaRRaR RaRaR [ R R
R]
= R (a R a) R = {0}
i.e., J is a nilpotent ideal of R.
But R has no non-zero nilpotent ideal, so we must have
J = {0} i.e., 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
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 14
RaRa+aRa+RaR+aRaR [ R2 R, RaaR
RaR]
= {0} [ R a R = a Ra
= 0]
i.e., 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 i.e. there exists a smallest
positive integer t such that at = 0.
Then, Bt = a .a .........a (t times) = at = {0}
i.e., 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.
Theorem 18 : Let R be a Noetherian ring. Then the sum of nilpotent ideals in R is again a
nilpotent ideal.
Proof : To prove the theorem, let us first prove that sum of finite number of nilpotent ideals is again
a nilpotent ideal.
For this, it is sufficient to prove that sum of two nilpotent ideals is again nilpotent because that result
will be true for any finite number of nilpotent ideals by induction. So, let A and B are two
nilpotent ideals, then there exist positive integers m and n such that Am = 0 and Bn = 0.
Let k = max (m , n) , then Ak = Bk = 0 i.e., product of any k elements of A or B is zero.
Now, we claim that
(A + B)2k = 0 i.e., product of any 2k elements of A + B is zero.
Let a1 + b1 , a2 + b2 ,..........., a2k + b2k A + B be any 2k elements where a1 A, b1 B.
Then,
(a1 + b1) (a2 + b2) ( a2k + b2k) = a1r1 b1s1 a2r2 b2s2 ..... a2r2kk b2s2kk
……(1)
where one out of ri and si (1 i 2k) is 0 and other is 1.
Now, A and B are ideals (if left) then
rai A and rbi b r R, 1 i 2k
In particular, any ai multiplied with some b j ’s on its left belongs to A
……(2)
and any bi multiplied with some a j ’s on its left belongs to B
……(3)
Now , if in any term of (1) , ai ’s appear greater than equal to k times, then that term can be made a
product of at least k elements of A , using (2), and hence that term becomes zero.
Similarly, if in any term of (1), bi ’s appear greater than or equal to k times, then that term can be
made a product of at least k elements of B, using (3), and hence that term becomes zero.
Therefore, by (1) , product of any 2k terms of A + B is zero i.e. A + B is a nilpotent ideal. Now by
induction it follows that finite sum of nilpotent ideals is nilpotent.
Now, we come to the main proof of the theorem, and let
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 15
J = I
be the sum of nilpotent ideals in R.
Because R is Noetherian, J is finitely generated ideal. Suppose J = < x1 , x2 ,........, xn >.
Then each xi lies in the sum of finitely many I ’s. i.e., there exist a finite subset i of such
that
xi I
i
n
Let =
i 1
i , then is a finite subset of and then we have
J = < x1 , x2 ,........, xn > I
J J= I
Hence J is sum of finite number of I ’s. But all I ’s are nilpotent, so by above proved result, their
sum J is also nilpotent.
Theorem 19 : In a Noetherian ring, every nil ideal is nilpotent.
Proof : Let R be a Noetherian ring and N be its nil ideal. We shall prove that N is nilpotent. Let
T be the sum of all nilpotent ideals of R. Now, we know that, in a Noetherian ring, sum of nilpotent
ideals is again a nilpotent ideal, so T must be a nilpotent ideal of R. Since R is Noetherian, so
R T is also a Noetherian ring . Now we claim that R T has no non-zero nilpotent ideals.
Let A T be a nilpotent ideal of R T , where A is an ideal of R and T A. Then, there exist
m
some positive integer m such that A T = {T} A m T = {T} Am T.
But since T is nilpotent, so there exists a positive integer k s.t. Tk = (0) , then
m k
A Tk = {0} Amk = {0}
A is nilpotent ideal. But T is the sum of all nilpotent ideals of R, so A T.
Hence, A = T A T = {T} i.e. A T is a zero ideal.
Therefore, R T has no non-zero nilpotent ideal. Now R T is a Noetherian ring and has no non-
zero nilpotent ideals, so R T has no non-zero nil ideals.
Now, N is a nil ideal of R, so N N T is also a nil ideal. We know that
N N T N+T T
which implies that N+T T is a nil ideal. But N + T is a ideal of R, so N+T T is a nil ideal of
R T.
But we have proved above that R T has no non-zero nil ideal, therefore
N+T T = {T}
N+T = T
N T
But T is a nilpotent ideal of R and N T. So N is also a nilpotent ideal of R.
Theorem 20 : Hilbert Basis Theorem : Let R be a Noetherian ring with unity. Then the polynomial
ring R[x] is also a Noetherian ring.
Proof : Let F and F be the families of left ideals of R and R[x] respectively. Let n be a non-
negative integer. We define a mapping n : F F by setting
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 16
a R : there exist a polynomial
n (I) = n n--1 {0} I F
ax bx ......... I , a 0
i.e., n (I) contains leading coefficients of all polynomials of degree n in I and 0.
We claim that n (I) is an ideal R i.e., n (I) F
Let a , b n (I) then there exist two polynomials say,
f (x) = axn + a1xn–1 +.......... I
and g(x) = bxn + b1xn–1 +……. I
But I is an ideal so f (x) g(x) I
(a b) xn + (a1 b1) xn–1+............ I
a b n (I)
Further, let r R ,
Then r f (x) I
raxn + ra1xn–1 +............ I
ra n (I)
Hence, n (I) is an ideal of I.
Now, we show that n (I) ni (I) i 0.
Let a n (I) then there exists a polynomial , say, f (x) = axn + a1xn–1 +............. I
Then xi f (x) I [ xi
R[x]]
a x n i a1 x n i –1 ........... I
a n i I i0
Therefore, n (I) ni (I) i 0
……(1)
Now, we claim that if I , J F with I J and n (I) = n (J) for all n 0 then I = J.
For this, it is sufficient to prove that J I. Let 0 f (x) J be a polynomial of degree m.
But m (I) = m (J), so there exist a polynomial, say gm(x) I with same leading coefficient as that of
f (x).
Now, f (x) gm(x) is either 0 or degree at most m 1.
If f (x) gm(x) = 0 , then
f (x) gm(x) I f (x) gm(x) + gm(x) = f (x) I
and if f (x) gm(x) 0 then
gm(x) I J, f (x) J f (x) gm(x) J
Because f (x) gm(x) J , we can similarly find gm–1(x) I such that f (x) gm(x) gm–1(x) J and
is either 0 or of degree at most m 2.
Continuing like this, we arrive, after at most m steps, at
f (x) gm(x) gm–1(x) ........... gi(x) = 0 where gm(x) , gm–1(x) ,......... I
But then f (x) I.
Hence J I and therefore, I = J and so our claim is established.
Now, let A1 A2 A3 ......... An ........... be an ascending sequence of left ideals of R[x].
Then for each non-negative integer n , n (A1) n (A2) n (A3) ......... is an ascending
sequence of left ideals of R. But R is Noetherian so this sequence of ideals must become stationary
after a finite number of terms i.e. there exists a positive integer k (n) such that
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-II PAGE 17
n A k ( n ) = n A k ( n ) 1 =
……(2)
Now, consider the collection { n (Ai) } n N , i N , of ideals of R, but since R is Noetherian, this
collection must have a maximal element, say, p (Aq).
Now, by (1)
p (Aq) n (Aq), for all n p
But p (Aq) is maximal, so we have
p (Aq) = n (Aq) for all n p
……(3)
Also, n (Aq) n (Aj) for all j q
……(4)
By (3) , (4) and using maximality of p (Aq) , we obtain
p (Aq) = n (Aq) = n (Aj) for all n p , j q
Therefore, we can choose, k (n) = q for all n p in (2).
Moreover, if we take s max k 1 , k 2 , ......, k p 1 , q then we must have
n A s n A s 1 n A s 2 for all n
Hence by the claim, which we proved above, we get
A s A s 1 A s 2
Hence, any sequence of left ideals of R[x] becomes stationary after a finite number of steps. Therefore
R[x] is Noetherian.
Important Notes
Note 1 : Throughout, unless otherwise stated, by a Noetherian (Artinian) ring we mean a left
Noetherian (Artinian) ring.
Note 2 : In view of the importance of Noetherian and Artinian rings in themselves, we rewrite two
theorems (theorem 1 and 2) for rings as follows:
Note 3 : Clearly every nilpotent ideal is nil. Since if A is nilpotent ideal then there exist a positive
integer n s.t. An = {0}. So for each a A, an An = {0} an = 0 , hence A is nil
ideal.
However, converse is not true in general.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 1
Def. Idempotent element : An element e in a ring R is called idempotent if e 2 e .
Def. Minimal ideal : A non-zero ideal M of a ring R is said to be minimal if any other non-zero
ideal of R is not properly contained in M.
OR
An ideal M of a ring is called minimal ideal if
(i) M {0}
(ii) for any ideal A M , either A = {0} or A = M.
Def. 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 i.e., R = R1 R2 .......... Rk , then
R op = R1op R 2op .......... R op
k
(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.
(v) Let M = M1 M2 ................ Mk be a direct sum of R-modules.
Hom R (M1 ,M1 ) Hom R (M 2 ,M1 ) ... Hom R (M k ,M1 )
Hom (M ,M ) Hom (M ,M ) ... Hom (M ,M )
Then HomR (M , M) R 1 2 R 2 2 R k 2
as rings.
... ... ... ...
Hom R (M1 ,M k ) Hom R (M 2 ,M k ) ... Hom R (M k ,M k )
Here the right hand side is a ring of k k matrices of the form
f11 f12 ... f1k
f f 22 ... f 2 k
11 , where f Hom (M , M )
ij R j i
... ... ... ...
f k 1 f k 2 ... f nk
Lemma : Let A be a minimal left ideal in a ring R. Then either A2 = {0} or A = Re , where e
is some non-zero idempotent element in R.
Proof : If A2 = {0} , then there is nothing to prove. So, let A2 {0} and we shall prove that
A = Re for some idempotent element in R. Now A2 {0} i.e. , A.A {0}, so there exists a
non-zero element
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 2
a A such that Aa {0} . Also Aa is a left ideal of R. But Aa A and minimality of A says
that A does not contain any non-zero left ideal, so we must have Aa = A.
Now, a A a Aa a = ea for some e A
Clearly e 0 because a 0.
Now, we shall prove that this element e is our required element i.e. , e2 = e and A = Re.
We see that e2a = [Link] = ea
(e2 – e)a = 0
Let B = {c A : ca = 0} , then B is a left ideal of R and clearly B A. Also, e.a = a 0
implies that e B, but e A , shows that B A. Again, using minimality of A, we must have B
= {0}.
Now, (e2 – e)a = 0 implies that e2 – e B and therefore e2 – e = 0 e2 = e
Further, e A and A is an left ideal of R, so Re A and Re {0} , because 0 e = e2 = e
e Re and then by minimality of A we must have A = Re,which completes the proof.
Theorem : Wedderburn Artin Theorem : Let R be a left artinian ring with unity and has no non-
zero nilpotent ideals. Then R is isomorphic to a direct sum of finite number of matrix rings over
division rings.
Proof: First of all , we claim that each non-zero left ideal in R contains at least one non-zero
idempotent element. So, let A be any non-zero left ideal of R.
If we consider the family of all non-zero left ideals of R which are contained in A, then this family
must have a minimal element , say M, because R is given to be Artinian. Now M is a non-zero
minimal left ideal of R and M A. By above lemma, we must have either M2 = {0} or M = Re
for some non-zero idempotent element e R.
Suppose M2 = {0} and consider the ideal B = MR, then
B2 = (MR)2 = MRMR MMR = M2R = {0}R = {0}
B2 = {0}
This shows that B is a nilpotent ideal of R but we are given that R has no non-zero nilpotent ideals,
so we obtain B = {0}. i.e., MR = {0}. But R is a ring with unity, so this gives M = {0} , which is
a contradiction as M was a non-zero ideal. Hence second possibility of M must occur i.e. M = Re
for some idempotent element e in R.
Now e = 1.e R.e = M A. Thus A contains a non-zero idempotent element. But A was any
non-zero left ideal, so we can conclude that every non-zero left ideal of R contains at least one non-
zero idempotent element.
Secondly, we claim that every non-zero left ideal of R is of the form Re for some non-zero
idempotent element e in R. So, let A be any non-zero left ideal of R.
Let us consider a family of left ideals , namely,
= {R(1 e) A : e is a non-zero idempotent element of A}.
As A contains at least one non-zero idempotent element by our first claim, so clearly is non-
empty. This non-empty family of left ideals of R must have a minimal element, because R is left
Artinian. Let this minimal element of be R(1 e) A. We claim that R(1 e) A = {0}.
Let if possible, R(1 e) A {0}. Then it is a non-zero left ideal of R and so by our first claim,
it must contain a non-zero idempotent element, say e1.
Now, e1 R(1 e) A implies that e1 R(1 e) so, e1 = r(1 e), for some r R
Now e1.e = r(1 e)e = r(e e2) = r(e e) = 0
……(1)
Let e = e + e1 e e1 then e A as e, e1 A.
Now, e.e = (e + e1 ee1) (e + e1 e e1)
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 3
= e2 + e e1 e2 e1 + e1e + e12 e1ee1 ee1e e e12 +ee1ee1
= e + ee1 ee1 +0+e1 0 0 ee1+0 [By (1) and e2 = e ,
2
e = e1]
1
= e + e1 ee1 = e
……(2)
and e1e = e1 (e + e1 ee1) = e1e + e12 e1ee1 = 0 + e1 0 = e1 0
……(3)
Let x R(1 e) be an arbitrary element , then x = r1(1 e) for some r1 R
and x = r1(1 e) = r1(1 e e1 ee1)
= r1(1 e) r1e1 r1ee1
= r1(1 e) r1(1 e)e1
= r1 (1 e) r1(1 e) r(1 e) [ e1 = r(1 e)]
= (r r(1 e)r) (1 e) R(1 e)
So, R(1 e) R(1 e) R(1 e) A R(1 e) A
……(4)
Now, by (2) and (3), it is clear that e is a non-zero idempotent element of A and so by definition
of ,we have R(1 e) A . But R(1 e) A is a minimal element of , so by (4), we
must have
R(1 e) A = R(1 e) A
Now we know that e1 R((1 e) A) e1 R(1 e) A
e1 R(1 e)
e1 = r(1 e), for some r R
e1e = r(1 e)e = r(e ee) = r(e e) = 0
which is a contradiction by (3) .
This establishes our claim that R(1 e) A = {0}.
Now, let a A be an arbitrary element.
Then a(1 e) R(1 e) and a(1 e) = a ae A [ a , e A]
a(1 e) R(1 e) A = {0}
a(1 e) = 0
a ae = 0
a = ae
A = Ae [ ‘a’ was an arbitrary element] Also , as e
A and Now A is a left ideal of R, so, re A r R Re A
Then, A = Ae Re A implies that A = Re
So, our second claim i.e., every non-zero left ideal in R is of the form Re for some non-zero
idempotent element in R, is established.
Let S be the sum of all minimal left ideal in R i.e. S = Ai , where A i i is the family of all
i
minimal left ideals in R. Clearly, S is a left ideal of R. Then we must have S = Re for some non-
zero idempotent element in R.
We claim that R(1 e) = {0}. Let, if possible, R(1 e) {0}, then there exists a minimal non-
zero ideal M contained in R(1 e). This ideal M is also contained in S = Re, since S is sum of
all minimal left ideals.
So, M Re R (1 e) = Re (R Re) = {0} M = {0}, which is a contradiction.
Hence, R(1 e) = 0 R Re = 0 R = Re.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 4
So, we obtain R = S R= A
i
i
Now, a minimal left ideal of R is a simple sub-module of R as a left module of R.
We know that “Let M be a R-module such that M = M i is the sum of a family of simple R-
i
modules Mi i , then there exists a sub-family Mi i such that M = M i ”
i
Using this result, there exists a sub-family A i i of the family of the minimal left ideals such that
R = A
i
i
Now 1 R and let 1 = ei1 ei 2 ....... ei n , where 0 ei j A i j and i j
R = R ei1 R ei 2 ........ R ei n
where each R ei j is a minimal left ideal as ei j A i j implies that R ei j Ai j
R ei j = Ai j because Ai j is a minimal left ideal.
After reindexing in above expression, we may write
R = R e1 R e2 ........ R en
In the family of minimal left ideals R e1, R e2 ,........, R en , choose a largest subfamily consisting of
all minimal left ideals that are not isomorphic to each other as left R-modules.
After renumbering, if necessary, let this subfamily be R e1 , R e2 ,........, R ek .
Suppose the number of left ideals in the family R e1, R e2 ,......., R en that are isomorphic to R ei in
ni.
Then
R R e1 .......... R e2 ........... ........ R ek ...........
n1 summands n2 summands nk summands
where each set of brackets contain pair wise isomorphic minimal left ideals, and no minimal left ideal
in any pair of brackets is isomorphic to a minimal left ideal in another pair.
We observe for R ei and R e j , i j , i.e., for two non-isomorphic minimal left ideals that
Hom R (R ei , R e j ) = {0}
Further , by Schur’s lemma , we know that
Hom R (R ei , R e j ) = Di , a division ring.
Now, making use of result (5), we obtain
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 5
D1 D1
....O
D D
1 1
n1 n1
D2 D2
D2 D2
Hom R (R,R) n2 n2
Dk Dk
O....
Dk Dk
nk nk
(D1 ) n1
(D 2 ) n2
= (D ) (D ) ....... (D )
1 n1 2 n2 k nk
(Dk ) nk
But we know the result that Hom R (R,R) R op as rings. So we obtain
R op (D1 )n1 (D2 )n2 ....... (Dk )nk
op op op
op op
R (D ) (D ) ........ (D )
1 n1 2 n2 k nk
R D D ........ D
op
1 n
1
op
2 n
2
op
k n
k
But we know that opposite ring of a division ring is a division ring and hence R is a finite direct sum
of matrix rings over division rings.
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 gi1 , where i denotes the complex conjugate of i . Then
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 6
n n
2
xx* i i gi
i 1 i 2
n
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 *r 0 . (We may assume r is even.) Set b aa * 2 . Then b2 0 and b b * . Thus, bb* 0 ,
r
which 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 Dni contains a copy K of F in its center. In this
i
way each Dni is a finite dimensional algebra over K (How?). Let D i : K n , and a D i . Then
i
2 n
1, a, a ,...., a 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 K F and completes the proof.
i
Example 1 : Let Dn be the n n matrix ring over a division ring D. Then Dn is an n2 – dimensional
vector space over D , and each left ideal as well as each right ideal of Dn is a subspace over D. Thus ,
any ascending or descending chain of left as well as right ideals can not contain more than n2 + 1
terms. Thus , Dn is both noetherian and artinian ring.
Theorem : Let R be a left or right artinian ring with unity and no non zero nilpotent ideals. Then R
is also right and left artinian.
Proof : Since R is artinian ring with unity so by Wedderburn – Artin theorem , it is isomorphic to a
finite direct sum of matrix rings over division rings. Because the matrix rings over the division rings
are both right and left artinians and noetherian and a finite direct sum of artinian and noetherian rings
is again a artinian and noetherian ring , we get that R is also a left and right notherian and artinian ring.
Example 2 : Let R = and A = show that A is an ideal of R and as a left R –
0 0 0 0
module it is simple and R/A is a field. Also prove that R is left artinian but not right artinian.
0
Solution : Let R = and A = 0 0 , then we prove that A is an ideal of R. Clearly ,
0 0
0 r m n
difference of any two matrices from A is again belonging to A. Now let A and 0 0 R
0 0
be any two elements , where m , n and r are rational numbers.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 7
0 r m n 0 0 m n 0 r 0 mr
Then 0 0 0 0 = 0 0 A and 0 0 0 0 = 0 0 A
Thus , A is an ideal of R.
0 r
Further , the mapping : A defined by r = is clearly an isomorphism , so
0 0
A . Since has no proper ideals. So A also has no proper ideals. Therefore , A as a left R –
module has no proper sub – modules and so A is simple as a left R – module. Thus A is left artinian.
0 a b a 0
Now , we define a mapping f : R setting f
0 0 0 0 0 0
Clearly f is an onto homomorphism , so by Fundamental theorem of homomorphism
0
R/Ker f
0 0
0 0
But Ker f = = A , so that R/A
0 0 0 0
Hence R/A is a field. Therefore , R/A as an R/A-module (or as an R - module) is artinian. Now , we
know that if submodule and quotient module of a module are left artinian , then module itself is also
left artinian. Using this result we obtain that R is left artinian.
Finally , we show that R is not right artinian. In fact there exists a strictly descending chain
0 2 0 2 2 0 2 3
0 0 …... of right ideals of R , which never becomes
0 0 0 0
stationary.
Def. Uniform module : A non – zero module M is called uniform if any two non – zero sub modules
of M have non zero intersection.
Def. Sub-isomorphic module : If U and V are uniform modules , we say U is sub-isomorphic to
V and write U V provided U and V contain non zero isomorphic sub modules.
Remark : Note is an equivalence relation [U] denotes the equivalence class of U.
Def. Primary module : A module M is called primary if each non – zero sub module of M has
uniform sub – module and any two uniform submodules of M are sub – isomorphic.
Example 3 : as a module is uniform and primary.
Theorem : Let M be a noetherian module or any module over a notherian ring R. Then each non –
zero sub module contains a uniform module.
Proof : Either M is noetherian or R is noetherian , so every sub modules of M is finitely
generated. For this reason , it is enough to show that xR contains a uniform submodule for any x( 0
) M. If M is notherian the sub module xR is also noetherian. But if R is noethrian , then xR
being homomorphic image of R , is notherian. Thus in both the cases , xR is noetherian. For
convenience , we shall call , a non – zero sub – module N of M large if N K {0} for all non –
zero submodules K of M.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 8
Consider now the family F of all sub – modules of xR which are not large. Clearly 0 F.
Since xR is noetherian , F has a maximal member say K. Because K is not large , K U = {0} for
some non – zero sub module U of xR. We claim U is uniform. Let if possible , U is not uniform.
Then, there exist sub – modules A and B of U such that A B = {0}. But then we shall prove that
(K A) B = {0}.
For this let x (K A) B be an arbitrary element then x K A and x B
x = k + a for some k K , a A and x = b for some b B
k+a = b
k = ba
b a K
But a , b U b a U and so k = b a K U = {0}.
Hence , k = 0 = b a a = b
Now a A and b B
So a = b A B = {0} , which implies that a = b = 0. Thus x = b = 0 and so (K A) B =
{0}
But this is a contradiction to the maximality of K. Hence our supposition is wrong. So , U is
uniform.
Hence each non – zero sub – module of M contains a uniform module.
Def. Ideal associated with a module : If R is commutative noetherian ring and P is a prime ideal
of R , then P is said to be associated with module M if R/P embeds in M.
or
If R is commutative noetherian ring and P is a prime ideal of R , then P is said to be associated
with module M if P = r(x) for some x M , where r(x) = {a R : xa = 0} denotes the annihilator
of x.
Def. P-primary module : A module M is called P – primary for some prime ideal P if P is the
only prime ideal associated with M.
Remark : If R is commutative noetherian ring and P is a prime ideal of R , then an R – module is
P – primary if and only if each non – zero sub – module of M is sub isomorphic to R/P.
Theorem : Let U be a uniform module over a commutative noetherian ring R. Then U contains a sub
module isomorphic to R/P for precisely one prime ideal P i.e. R/P embeds in U for exactly one prime
ideal P.
Proof : Consider the family F of annihilator ideals r(x) , where 0 x U i.e., F = {r(x) : x
( 0) U}
Since R is noetherian , there exists a maximal member say r(x) in F.
We claim that P = r(x) is prime ideal.
Let ab r(x) such that a r(x) we prove that b r(x). Since ar(x) , xa 0 and so r(xa) F.
Now if y r(x) be any element , then xy = 0 which implies xay = xya = 0.a = 0 and so y r(xa).
Hence r(x) r(xa).
By maximality of r(x) , we obtain
r(x) = r(xa)
……(1)
Now ab r(x) [Link] = 0
b r(xa) b r(x) [By (1)]
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 9
This proves that P is a prime ideal.
Now , we define a mapping : xR R/P = R/r(x) by setting (xa) = a + r(x)
Let us prove that is an isomorphism
(i) is a homomorphism : (xa + xb) = (x(a + b)) = (a + b) + r(x)
= (a + r(x)) + b(r(x))
= (xa) + (xb).
and r ( xa) xra ra r ( x) r a r ( x ) r ( xa )
(ii) is one – one : Let (xa) = (xb)
a + r(x) = b + r(x)
(a b) + r(x) = r(x)
a b r(x)
x(a b) = 0
xa xb = 0
xa = xb.
(iii) is onto : Let a + r(x) R/r(x) be arbitrary element, then xa xR and (xa) = a + r(x)
Hence xR R/P as xR is sub – module of U, so we have proved upto now that R/P embeds in U
for one prime ideal P i.e. U contains a sub module xR isomorphic to R/P for one prime ideal P.
We shall show that it happens for exactly one prime ideal P. Let Q be
any other prime ideal such that R/Q is embeddable in U. Then R/P and R/Q must be sub –
isomorphic to each other and our aim is to prove that they are isomorphic. Since R/P and R/Q are sub
isomorphic, there exist cyclic sub modules xR of R/P and yR of R/Q such that xR yR .
But xR R/P and yR R/Q so R/P R/Q , which gives that P = Q. Hence R/P is embeddable in
U for exactly one prime ideal P. This completes the proof.
Theorem : Let M be a non – zero finitely generated module over a commutative noetherian ring R.
Then there are only a finite number of primes associated with M.
Proof : Consider the family F consisting of the direct sums of cyclic uniform sub – modules of M.
Since R is commutative noetherian ring it must have atleast one uniform cyclic sub – module so that
F is non – empty. Let us define a partial order on F by
xi R y j R iff I J and xiR yiR for i I
iI jJ
By Zorn’s lemma , F has a maximal member K = x j R , say.
jJ
Since M is finitely generated over commutative noetherian ring , K must be finitely generated and
let
l
K = x jR
j 1
Then , there exist xj aj xjR such that r(xj aj)= Pj , the prime ideal associated with xjR.
Then we have P1 , P2 ,…..,Pl are prime ideals associated with x1R , x2R ,…….., xlR and hence with
M.
l
Now set xj x j j and K = xj R
j 1
Let Q = r(x) be any associated prime ideal of M. We shall prove that Q = Pj for some j , 1 j l .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-III PAGE 10
This will prove that there are only a finite number of prime ideals associated with M. Since K is a
maximal member of F , K as well as K have the property that each intersects non – trivially with
any non – zero sub module L of M.
l
Now , let 0 y xR K be any element, then y xR and y K = xj R
j 1
l
y = xb for some b R and y= x b
j 1
j j , where b j R
We claim that r ( xj b j ) r ( xj ) whenever ( xj b j ) 0
Clearly, r ( xj ) r ( xj b j ) .
On the other hand , let c r ( xj b j )
xj b j c = 0 bj c r ( xj ) = Pj
But xj b j 0 bj r ( xj ) = Pj
As Pj is a prime ideal , we must have c Pj = r ( xj ) . Hence r ( xj b j ) r ( xj )
So , we obtain r ( xj ) = r ( xj b j )
By this argument , we obtain r(x) = r(xb)
Now , Q = r(x) = r(xb)
l
= r(y) = r xjb j
j 1
l
= r x b = r x
j 1
j j
j
j = P
j
j
omitting those Pj’s which arises from xj b j = 0 and so is a subset of {1 , 2 ,….., l}
Hence , Q Pj for all j
Also , P
j
j P
j
j = Q
Since Q is a prime ideal , atleast one Pj appearing in the product P j
j must be contained in Q.
Hence Q = Pj for some j.
Theorem : Noether – Lasker Theorem : Let M be a finitely generated module over a commutative
noetherian ring R. Then there exists a finite family N1 , …..Nl of sub modules of M such that
l l
(i) Ni = 0 and
i 1
N i 0 for all 1 i0 l
i 1
i i0
(ii) Each quotient M/Ni is a Pi – primary module for some prime ideal Pi.
(iii) The Pi are all distinct 1 i l .
(iv) The primary component Ni is unique iff Pi does not contain Pj for any j i.
Proof : First reproduce the above theorem. Let {Ui} , 1 i l be uniform sub modules obtained in
above theorem. Choose Ni to be a maximal member in the family {K M ; K contains no sub –
module sub isomorphic to Ui}. With this choice of N1 ,……, Nl ; (i) , (ii) , (iii) and (iv) follows
directly.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 1
Def. Linear Transformation : Let V1 and V2 be two vector spaces over the same field F.
A mapping T from V1 to V2 is called a linear transformation or a vector space
homomorphism if it satisfies the following properties :
(i) T x y T x T y x , y V1
(ii) T x T x x V1 and F
Results :
1. Let V1 and V2 be two vector spaces over the same field F. A mapping T : V1 V2 is
a linear transformation if and only if T x y T x T y , F , x , y V1 .
2. Let T : V1 V2 be a linear transformation. If u1, u2,......., un are L. D. vectors in V1 ,
then T u1 ,T u2 ,........,T un are L.D. vectors in V2 .
3. A linear transformation T : V1 V2 is one – one if and only if it maps every linearly
independent subset of V1 into a linearly independent subset of V2 .
Def. Kernel of A L.T. : For any linear transformation T : V1 V2 , the kernel of T (ker T)
is defined as Ker T {v V1 : T v 0} . Ker T is also called the null space of T and
sometimes denoted by N(T).
4. Let T : V1 V2 be a L. T. Then
(i) Ker T is a subspace of V1.
(ii) For any subspace W of V1 , T(W) is a subspace of V2 .
Def. Quotient space : Let W be a subspace of a vector space V over a field F. Then the
set V W x W : x V is a vector space over F where addition and scalar
multiplication are defined as x W y W x y W ,x W , y W V /W
and x W x W , F .
This vector space V W is called quotient space.
5. If V and W are subspaces of a finite dimensional vector space V over F, then
dim V W dimV dimW (V W ) .
6. If V 0 , the null set is defined to be basis of V, so that dim V O .
7. If V W 0 in the Result –(1), then, we have dim V W dimV dimW
8. dim V /W dimV dimW .
9. For any subspace W of finite – dimensional vector space V, dim W dim V.
Further W = V if and only if dimV dimW .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 2
Def. Let U and V be two vector spaces over the same field F. The set of all linear
transformations of U into V is denoted by HomF (U , V) or simply by Hom (U , V).
Result : Let U and V be two vector spaces over the same field F. Then Hom (U , V)
is a vector space over F where addition and scalar multiplication in Hom (U , V) are
defined as
T T x T x T x ,TT Hom U ,V
1 2 1 2 1 2
and T x T x , T Hom (U , V) x U .
Def. Algebra : A ring R is said to be an algebra over a field F if the additive group
R , of R is a vector space over F and for any x, y R , F , we have
xy x y x y .
Theorem : For any vector space V(F) , Hom (V , V) is an algebra over F.
Proof : We know that Hom (V , V) is a vector space over F. For any
T1, T2 Hom U ,V , the Composite T1T2 :V V is such that for any v1, v2 V and
, F
T T (v v ) T [T (v v ) T T v T v ]
1 2 1 2 1 2 1 2 1 2 1 2 2
T T v T T v TT v TT v
1 2 1 1 2 2 1 2 1 1 2 2
This proves that TT
1 2
is a linear transformation on V so that T1T2 Hom V ,V .
Now for T1, T2, T3 Hom V ,V , v V , F
T T T v T [(T T )v ] T T v T v
1 2 3 1 2 3 1 2 3
TT1 2
v TT1 3
v TT1 2 1 3
TT v
Therefore, T1 T2 T3 TT
1 2
TT
1 3
Similarly, T 2
T3 T1 T2T1 T3T1
Since composition of mappings is always associative, so we get Hom (V , V) is a ring.
Now, [T1(T2 ) v T1 (T2 ) v T1 T2 v ] T1 {T2 v }
Hence , we get , T1(T2 ) (T1)T2 TT 1 2
Therefore, Hom (V , V) is an algebra over F.
Notation : Hom (V , V) will also be denoted by A(V).
Result : Let U and V be two vector spaces over a field F, of dimensions m and n
respectively. Then Hom (U , V) is a vector space over F of dimension mn.
Result : For any vector space V of dimension n, A(V) is an algebra of dimension n2 over F.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 3
Def. Let g x 0 1x ... n x n F x be a polynomial and T A V be any linear
transformation. We say that T is a root of g(x) if g T 0 1x ... nT n 0 zero
linear transformation on V.
Theorem 1 : If V is an n dimensional vector space over F, then, given any element T
in A(V), there exists a non trivial polynomial q x F x of degree at most n 2 such that
q T 0 .
2
Proof : Since A(V) is an algebra of dimension of n 2 , so I ,T , T 2,.T n are linearly
dependent over F. So there exist scllars i' ’s 0 i n 2 in F, not all zero , such that
2
0 I 1T 2T 2 n 2T n 0 …….(1)
0 1x 2x 2 . n 2x n F x , then
2
Now by (1) , T clearly satisfies
q x F x , which is of degree at most n2.
Def. Minimal polynomial : Let 0 T A V . A monic polynomial of lowest degree of
which T is a root is called minimal polynomial of T.
Remark : By above theorem, we observe that degree of minimal polynomial can not exceed
n2
Def. Right invertible : A linear transformation T T V is called right invertible if there
exist S A V such that TS I .
Def. Left invertible : A linear transformation T A V is called left invertible if there exist
U A V such that UT I .
Def. Invertible linear transformation : A linear transformation T A V is called invertible
or regular if T is both right and left invertible.
Def. Singular linear transformation : An element in A(V) which is not regular is called
singular.
Theorem 2 : If T A V is regular then its right and left inverses are same i.e. there exist
S A V such that ST TS I
Proof : Since T is both right and left invertible there exist S , S A V such that
TS I and S T I .
Now, S ' IS ' S "T S ' S " TS ' S " I S " .
Putting S S " S , we get ST TS I .
Remark : If T A V is regular then there exist a unique S s.t. TS ST I .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 4
Let S A V be another element s.t. TS ' S 'T I , then we have
S IS S 'T S S ' TS S ' , which proves the inqueness of S.
Such an S A V is called inverse of T and we denote it by T –1 .
Theorem 3 : If V is finite - dimensional over F , Then T A V is invertible iff the
constant term of the minimal polynomial of T is non – zero.
Proof : Let m x x k k 1x k 1 ..... 1x 0 be the minimal polynomial of T. Then
T k k 1 T k 1 ...... 2T 2 1T 0I 0
T T k 1 k 1T k 2 ....... 2T 1I
T k 1
k 1T k 2 ....... T I T I
2 1 0
.........(1)
Now, suppose T is regular. We shall prove that constant term i.e 0 0 .
Let, if possible, 0 0 . Then by (1)
T k 1 k 1T k 2 ....... 1I 0 [Since T 0 as T is regular]
which shows that T satisfies a non – zero polynomial over F of degree less than k.
This contradicts the fact that degree of minimal polynomial is k. Hence 0 0 .
Conversely, let us suppose that 0 0 . Then by (1)
TS ST I , where
S 01 T k 1 k 1T k 2 ....... 1I .
Hence T is regular.
Corollary 1 : If V is finite dimensional over F, and if T A V is invertible then T 1 is
a polynomial expression in T over F.
Proof : By above theorem,
T 1 S 01 T k 1 k 1T k 2 ....... 1I which is a polynomial in T over F.
Corollary 2 : If V is finite dimensional over F and if T A V is singular, then there
exist an S 0 in A(V) such that ST TS 0
Proof : Because T is not regular, the constant term of its minimal polynomial must be
0. That is
m x x k k 1x k 1 ..... 1x , so we must have ,
T k k 1T k 1 ..... 1T 0
T T k 1 k 1T k 2 ....... 1I 0
If S T k 1 k 1T k 2 ....... 1I , then S 0 , since x k 1 k 1x k 2 ..... 1 is of lower
degree than p x . and we obtain
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 5
ST TS 0
Corollary 3 : If V is finite dimensional over F and if T A V is right invertible, then
it is invertible.
Proof : Let TU I . If T were singular, there would be an S 0 such that S T 0
(By Cor 2).
But then, O OU ST U S TU SI S 0 , a contradiction. Thus T is regular.
Theorem 4 : A linear transformation T A(V) is singular iff there exist v ( 0) s.t T
v = 0.
Proof : We know that if T A(V) is singular then there exist S ( 0) A(V) s.t. T S =
ST =0 (Cor-2). Since S 0 , there exist an element w V s.t. S(w) 0. Let S(w)
= v , then this v is our required element. We note that T(v) = T(S(w)) = (T S)
(w) = O (w) = 0
Conversely : Let there exist v( 0) V such that Tv = 0. Let, if possible, T is not
singular. Then T must be regular. So there exist an S A(V) such that T S = S T
= I. Then
ST (v ) S (T (v )) S (0) 0 and again ST (v ) I (v ) v
From these two, v 0 , a contradiction. Hence T must be singular.
Remark : Throughout this unit, unless otherwise stated, V will be finite dimensional vector
space.
Theorem 5 : Let T( 0) A(V) and S A(V) be regular. Then T and S 1T S have same
minimal polynomials.
Proof : For any g(x ) 0 1x k x k F [x ] , we have
2 k
g S 1T S 0.I 1 S 1T S
2 S 1T S
.. k S 1TS …….(1)
But we see that
2
S 1
T S S 1T S S 1T S S 1T 2S ,
3
S 1
T S S 1
T S S 1T S S 1T S S 1T 3S
i
In general, S 1T S S 1T iS for i 0 .
Now, g S 1T S 0.I 1S 1T S 2S 1T 2S .... kS 1T kS
S 1( 0I 1T 2T 2 . kT k ) S
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 6
S
S 1 g T
Hence g T 0 if and only if g S 1T S 0 ………(2)
Now, let m x and m ' x are the minimal polynomial of T and S 1T S respectively.
Then, we have [using (2)],
m T 0 ,
m S 1T S 0
m ' T 0 , m ' S T S 0
1
Then definition of minimal polynomial gives
m(x ) m (x ) and m (x ) m (x )
But since m x and m (x ) are monic, so
m x m' x i.e. minimal polynomials of T and S 1T S are same.
Def. Characteristic root : An element F is said to be a characteristic root or
characteristic value of T A(V) if there exist some v 0 in V such that T(v) = v
Def. Characteristic vector : An element 0 v V is said to be a characteristic vector of
T A(V) if there exist some F s.t. T(v) = v and then we say that v is characteristic
vector of T belonging to characteristic root .
Remark : Characteristic vector and characteristic root are also known as eigen vector and
eigen root respectively.
Theorem 1 : If v is non-zero eigne vector of T A(V) then F s.t. T(v) = v is
unique.
Proof : Let T(v) = v and T(v) = v for some , F. Then,
( )v = 0 = 0 since v 0
i.e. = which proves uniqueness of .
Theorem 2 : For any eigen value of T, the set of all vectors v V s.t. Tv = v forms
a non-zero subspace of V.
Proof : Let W = {v V : Tv = v}. By definition of eigen value, there exists
v ( 0) V s.t. Tv = v. Thus v W and hence W {0}.
Now, let u , v W and , F. Then we have T(u) = u and T(v) = v
Consider T( u + v) = T(u) + T(v) = u + v = ( u + v)
u + v W and hence W is a non-zero subspace of V.
This subspace W of V is called eigen space belonging to .
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 7
Theorem 3 : If F is a characteristic root of T A(V) , then for any polynomial
q(x) in F [x] , q( ) is a characteristic root of q(T).
Proof : Now F is a characteristic root of T , so by definition, there is a non-zero
vector v V s.t. T(v) = v.
Now , T2(v) = T(T(v)) = T( v) = T(v) = 2v
Continuing in this manner, we obtain, Tk(v) = k for all positive integers k.
If q(x) = 0 + 1x + 2x2 +………+ mxm , then q(T) = 0I + 1T + 2T2 +………+
mTm
and hence (q(T)) (v) = ( 0I + 1T + 2T2 +..............+ mTm )(v)
= 0I (v) + 1T (v) + 2T2 (v) +...............+ mTm (v)
= 0 v + 1 v + 2 2 v +……….+ m Tm (v)
= ( 0 + 1 + 2 2
+………+ m m
) (v)
= (q( )) v
Hence , q( ) is a characteristic root of q(T).
Theorem 4 : If F is a characteristic root of T A(V) then is a root of minimal
polynomial of T. In particular, T has only a finite number of characteristic roots in F.
Proof : Let p(x) be the minimal polynomial of T over F , thus p(T) = 0.
If F is a characteristic root of T, there is a v 0 in V such that T(v) = v.
Let p(x) = 0 + 1x + …….+ mxm , then by the same calculations as that of above
theorem, we obtain, (p(T)) (v) = (p( )) v
But p(T) = 0 , so by property of vector space, we must have p ( ) = 0.
Therefore, is a root of p(x).
Now, p(x) has only finite number of roots in F, because deg p(x) n2 where n = dim
VF , therefore there can be only a finite number of characteristic roots of T in F.
Theorem 5 : If 1 , 2 ,……, k in F are distinct characteristic roots of T A(V) and
if v1 , v2 ,………,vk are characteristic vectors of T belonging to 1 , 2 ,………, k
respectively, then v1 , v2 ,………,vk are linearly independent vectors over F.
Proof : Let, if possible , v1 , v2 ,………,vk are linearly dependent over F, then there is a
relation of the form 1v1 + 2v2 +…………+ kvk = 0 where 1 , 2 ,……., k are all in
F and not all zero. In all such relations, there is one relation having as few non-zero
coefficients as possible. By suitably renumbering the vectors, we can assume this shortest
relation to be
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 8
1v1 + 2v2 +………..+ jvj = 0 , 1 0, 2 0,......., j 0 ……(1)
Now we know that, T(vi) = i vi , so applying T to equation (1), we obtain
T ( 1v1 + 2v2 +………….+ jvj) = T(0)
1T(v1) + 2 T(v2) +………….+ jT(vj) = 0
1 1v1 + 2 2v2 +…………+ j jvj = 0 …….(2)
Multiplying equation (1) by 1 and subtracting from equation (2), we obtain
2
1 2v 2 3 1 3v3 ..... j 1 jv j 0 ……..(3)
But since ’s are distinct, so 2 1 0, 3 1 0,....., j 1 0
Also, 2 0 , 3 0 ,………., j 0 .
Hence i 1 i 0 for 2 i j
Thus we have produced relation (3) which is shorter than relation (1) , a contradiction.
Hence, v1 , v2 ,………,vk must be L. I.
Cor 1 : If A(V) and dim VF = n then T can have at most n distinct characteristic roots
in F.
Proof : By above theorem , it is clear that any set of distinct characteristic roots of T gives
a corresponding set of linearly independent characteristic vectors and we know that any set
of linearly independent vectors in V can have at most n elements. Hence, T can have
at most n distinct characteristic roots in F.
Cor 2 : If T A(V) and if dim VF = n , and if T has n distinct characteristic roots
in F, then there is a basis of V over F which consists of characteristic vectors of T.
Proof : Let 1 , 2 ,……., n are n distinct characteristic roots of T in F , then there
exist corresponding characteristic vectors v1 , v2 ,……,vn in V such that Tvi = i vi ,
1 i n.
Now, we know that any n linearly independent vectors in a vector space of dimension n
from a basis , so {v1 , v2 ,……..,vn} must be a basis of VF.
Def. Matrix of a linear transformation : Let V be a n – dimensional vector space over a
field F and let v1 , v2 ,………,vn be a basis of V over F. If T A(V), then T(v1) ,
T(v2),…….,T(vn) must all be inV. Each of these can be expressed in a unique way as
linear combination of
v1 , v2 ,………,vn.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 9
Thus, T(v1) = 11 v1 + 12 v2 +………..+ 1n vn
T(v2) = 21 v1 + 22 v2 +……….+ 2n vn
T(vn) = n1v1 + n2v2 +……..…+ nnvn
Then the matrix if T in the basis v1 , v2 ,……..,vn written as m(T) is
11 12 ... 1n
22 ... 2n
m (T) = 21
... ... ... ...
n 1 n2
... nm
Another Form of Cor 2 : If T A(V) and if dim VF = n, and if T has n distinct
characteristic roots in F, then there is a basis of V over F in which the matrix of T is
diagonal.
Proof : If 1 , 2 , ……….., n are distinct characteristic roots, then by COR-2, there
exist a basis {v1 , v2 ,……..,vn} of V over F such that T(v1) = i vi
i.e. T(v1) = 1v1 + 0.v2 +……….+ [Link]
T(v2) = 0.v1 + 2v2 +………+ [Link]
T(vn) = 0.v1 + 0.v2 +…………+ nvn
0 ... 0
1
0 ... 0
Then, matrix of T in the basis {v1 , v2 ,………,vn} is m(T) = 2
... ... ... ...
0 0 ... n
Example : Let V be the vector space of all polynomials over F the field of real numbers,
of degree less than or equal to 3. Let D be the differentiation operator defined by
D( 0 + 1x + 2x2 + 3x3) = 1 + 2 2x + 3 3x2 .
Let us compute the matrix of D in the basis or relative to the basis
v1 = 1 , v2 = x , v3 = x2 , v4 = x 3 . Now
Dv1 = 0 = 0.v1 + 0.v2 + 0.v3 + 0.v4
Dv2 = 1 = 1.v1 + 0.v2 + 0.v3 + 0.v4
Dv3 = 2x = 0.v1 + 2.v2 + 0.v3 + 0.v4
Dv4 = 3x2 = 0.v1 + 0.v2 + 3.v3 + 0.v4
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 10
By the definition of matrix of linear transformation in a given basis, we see the matrix of
0 0 0 0
1 0 0 0
D in the basis v1, v 2, v3, v 4 is m1(D) =
0 2 0 0
0 0 3 0
Now, again, let us compute matrix of D in the basis
u1 = 1 , u2 = 1+x , u3 = 1+x2 , u4 = 1+x3
Now, Du1 0 0. u1 0. u2 0. u 3 0. u 4
Du2 1 1. u1 0. u2 0. u3 0. u4
Du3 2x 2. u1 2. u2 0. u3 0. u4
Du 4 3x 2 3. u1 0. u2 3. u 3 0. u 4
0 0 0 0
1 0 0 0
Then, the matrix m2 (D) of D in the basis u1, u2, u 3, u 4 is m2 (D) =
–2 2 0 0
–3 0 3 0
Hence, we see that matrices of D depends completely on the basis. Although different from
each other , they still represent the same linear transformation D.
Remark : Since the basis used at any time is completely at our choice , so given a L. T. T
we shall be seeking a basis in which the matrix of T has a particularly nice form. By nice
form we mean a matrix which mostly resembles a diagonal matrix. Now we shall discuss the
problems of choosing such basis of V for a given T A(v). In this course we shall be
identifying a linear transformation with its matrix in some basis of V.
Result : Let V be a n-dimensional vector space over F and v1 , v2 ,………,vn is a basis of
V over F, let S , T A(V) having matrices A and B respectively in the given basis i.e.
m(S) = A and m(T) = B. Then in this basis we have
(i) S = T iff A = B (ii) m (S + T) = A + B
(iii) m( S) = A (iv) m (S T) = A B
Remark : Let F be a field and Fn be the set of n n matrices over F. Then Fn is an
algebra of matrices. Let V be an n-dimensional vector space over F and let v1 , v2 ,……,vn
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 11
be a basis of V over F. Let T A(V), then the matrix m (T) in the basis v1 , v2 ,……,vn
is an element of Fn. In fact, for every T A(V) there is corresponding matrix in Fn and
for every matrix of Fn there is a linear transformation T A(V). the mapping from A(v)
into Fn defined by mapping T onto m (T) is an algebra isomorphism, of A(V) onto Fn.
In light of the very specific nature of the isomorphism between A(V) and Fn, we shall often
identify a linear transformation with its matrix in some basis and A(V) with Fn. Further, T
A(V) is invertible iff m(T) is invertible in Fn.
Theorem 6 : The element F is a characteristic root of T A(V) iff I T is
singular.
Proof : If is a characteristic root of T then there exist a non – zero vector v V such
that
T v = v ( I T) (v) = 0
So, by a previous theorem, I T must be singular.
Conversely : Suppose I T is singular. Then there exist a v ( 0) in V such that
( I T) (v) = 0 I (v) T (v) = 0 T(v) = v.
Theorem 7 : If T A(V) and if p(x) is the minimal polynomial for T over F, suppose that
p(x) has all its roots in F. Prove that every root of p(x) is a characteristic root of T.
Proof : Let p(x) F [x] be minimal polynomial of T over F and deg p(x) = k and let
1 , 2 ,.…..., k are roots of p(x) such that i F , 1 i k.
We shall prove that i (1 i k) is characteristic root of T.
Now, p(x) = (x 1) (x 2)…….(x k) and p(T) = 0, so
( 1 I T) ( 2 I T)……….( k I T) = 0 …….(1)
Let, if possible, i is not a characteristic root of T. Then, by above theorem, i I T is not
singular i.e. i I T is regular. So there exist S A(V) such that
S ( i I T) = ( i I T) S = I .…..(2)
Now, since ( i I T) ( j I T) = ( j I T) ( i I T) , so we can rewrite equation (1) as
( i I T) ( 1 I T)……….( i –1I T) ( i +1 I T)………( k I T) = 0
Pre – multiplying this equation by S and using (2), we obtain,
( 1 I T)….........( i–1 I T) ( i+1 I T)...........( k I T) = 0 …….(3)
Now, let q(x) = (x 1)……(x i–1) (x i+1)………(x k) then clearly q(x) F [x] and
using (3), we also have q(T) = 0.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 12
But p(x) is minimal polynomial, so it will divide any other polynomial satisfied by T. So
p(x ) q (x ) which is not possible since deg q(x) = k 1. Hence i must be a characteristic root
of T.
Def. A subspace W of V is said to be in variant under T A(V) if
T (W) W i.e. T(w) W w W.
Lemma : If W V is invariant under T, then T induces a linear transformation T on
V W , defined by T (v + W) = Tv + W. If T satisfies the polynomial q(x) F[x], then
so does T . If p1(x) is the minimal polynomial for T over F and if p(x) is that for T, then
p1(x ) p(x ) .
Proof : Let us denote V W by V and an element v + W of V W by v . Let T on
V be defined as T v = T (v + W) = T(v) + W.
(i) T is well – defined : Let v1 = v1 +W and v 2 = v2 +W be two elements of V s.t.
v1 = v2 v1 +W = v2 +W v1 v2 W
But W is invariant under T, so T(v1 v2) W T(v1) T(v2) W
T(v1) +W = T(v2) +W T ( v1 ) = T ( v 2 ).
(ii) T is linear transformation : We observe that
T ( v1 + v 2 ) = T ((v1+ W) + (v2 + W)
= T (v1 + v2 + W)
= T (v1 + v2) + W [By def. of T ]
= T(v1) + T(v2) + W [ T is a L. T.]
= T(v1) + W + T(v2) + W
= T ( v1 ) + T ( v 2 ).
and T ( v ) = T ( (v+W)) = T ( v + W) = T( v) + W
= T(v) + W = (T(v) + W) = T (v +W) = T ( v ).
Now, if v = v + W V W = V , then, we note that
( T )2 ( v ) = T ( T (v + W)) = T (T(v) + W) = T(T(v) + W
= T2(v) + W = T2 (v +W) = T2 ( v )
Continuing in this fashion, we obtain
T
k
Tk = for all k 0 ……..(1)
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 13
Now, let q(x) = 0 + 1x + …….+ mxm F [x] be any polynomial. Then, we observe
that
q T = 0 I + 1 T + 2 T +………+ m T
2 m
= 0 I + 1 T + 2 T2 +………+ m Tm [By (1)]
= 0 I 1 T 2 T 2 ........ m T m
= q(T)
So, if q(T) = 0 then q T = 0 = zero transformation on V .
Hence, if T satisfies q (x) then so does T . Now, let p(x) and p1(x) be the minimal
polynomials of T and T respectively. So, we have,
p (T) = 0 and p1 T = 0.
But we have proved above that if T satisfies any polynomial over F then T also satisfies
that polynomial so p T = 0. Since p1(x) is minimal polynomial of T , so we must have
p1(x ) p(x ) .
Remark : (i) A square matrix is said to be lower triangular if all the entries above the main
diagonal are zero and said to be upper triangular if all the entries below the main diagonal
are zero. Further, a matrix is said to be triangular if either it is lower triangular or it is
upper triangular.
(ii) If all the characteristic roots of T A(V) are in F then all the roots of minimal
polynomial of T over F lie in F.
Theorem 1 : If T A(V) has all its characteristic roots in F, then there is a basis of V in
which the matrix of T is triangular.
Proof : We shall prove the theorem by induction on dimension of V over F.
Let dim V = 1 and F be a characteristic root of T in F. Then there exist a non
– zero vector v0 V such that T (v0) = .v0.
Since v0 is a non – zero vector and dim V = 1, so {v0} is a basis of T. Hence matrix of
T in this basis is [ ]1 1 which is triangular. Hence theorem is true for dim V = 1.
As our induction hypothesis, we assume that the theorem is true for all vector spaces over F
of dimension n 1 and let V be of dimension n over F.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 14
The linear transformation T on V has all its characteristic roots in F and let 1 F be a
characteristic root of T. Then these exist v1 ( 0) V such that Tv1 = 1v1
LetW = < v1 > = { v1 : F}, then W is a one dimensional subspace of V. Let w W
be any element, then
w = 1 v1 for some 1 F
and T(w) = T( 1v1) = 1 T(v1) = 1 1v1 W implies that W is invariant
under T.
Let V = V W then dim V = dim V dim W = n 1.
Now, by above lemma, T induces a linear transformation T : V W V W defined
by
T (v + W) = T (v) +W
such that minimal polynomial of T over F divides the minimal polynomial of T over F.
Now, since all the roots of minimal polynomial of T over F lie in F and minimal
polynomial of T divides minimal polynomial of T over F, so all the roots of minimal
polynomial of T are in F and therefore all the characteristic roots of T are in F.
Now, dim V = n 1 < n and T : V V s.t. all the characteristic roots of T are
in F, hence by our induction hypothesis, there is a basis v 2 , v 3,………, v n of V over
F s.t.
T ( v 2) = 22 v 2
T ( v 3) = 32 v 2 + 33 v 3
T ( v i) = i2 v 2 + i3 v 3 +……….+ ii v i
T ( v n) = n2 v 2 + n3 v 3 +……………………+ nn v n
Now, we know by the theory of vector spaces that
“ If x , x , .., x is a basis of W over F and y W , y W ,, y
1 2 m 1 2 n
W is a basis
of V W over F then x , x , , x , y , y , , y is a basis of V over F ”.
1 2 m 1 2 n
Using this result, we conclude that { v1 , v2 ,…….,vn} is a basis of V over F. We shall prove
that this basis is our required basis i.e. matrix of T in this basis is triangular.
Now, T ( v 2) = 22 v 2 T (v2) + W = 22 v2 + W
T(v2) 22 v2 W = < v1 >
T(v2) 22 v2 = 21 v1 for some 21 F.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 15
T(v2) = 21 v1 + 22 v2
Similarly, T ( v i) = i2 v 2 + i3 v 3 +………….+ ii v i
T (vi) + W = i2 v2 + i3 v3 +................+ ii vi +W
T (vi) i2 v2 i3 v3 ............... ii vi W = < v1 >
T (vi) i2 v2 i3 v3 ………. ii vi = ii vi for some ii F
So we have obtained
T (v1) = 1 v1 = 11 v1 + 0. v2 +…….+ 0. vn
T (v2) = 21 v1 + 22 .v2 +………+ 0. vn
.........................................
T (vn) = n1 v1 + n2 v2 +……….+ nn vn
and so matrix of linear transformation T in the basis { v1 , v2 ,……., vn} is
11 0 ... 0
22 ... 0
m(T) = 21
... ... ... ...
n 2 n 2 ... nn nn
which is triangular matrix. So we have produced a basis of V in which the matrix of T is
triangular. This completes the proof.
Theorem 2 : If V is n-dimensional over F and if T A(V) has all its characteristic roots
in F, then T satisfies a polynomial of degree n over F.
Proof : Since T has all its characteristic roots in F, so by above theorem, we know that
there is a basis { v1 , v2 ,……., vn} of V over F such that
T v1 = 11 v1
T v2 = 21 v1 + 22 v2
T vi = i1 v1 + i2 v2 +.......+ ii vi
T vn = n1 v1 + n2 v2 +……………...+ nn vn
Putting ii = vi in above expressions and transposing the last terms on right hand side, to
the left hand side, we obtain
(T 1 I) (v1) = 0
(T 2 I) (v2) = 21 v1
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 16
(T i I) (vi) = i1 v1 + i2 v2 +............+ i,i–1 vi–1
(T n I) (vn) = n1 v1 + n2 v2 +………+ n,n –1 vn –1 …….. (1)
We have,
(T 1 I) (v1) = 0 ………(2)
Now , (T 1 I) (T 2 I) (v1) = (T 2 I) (T 1 I) (v) [ (T 1 I) and (T 2 I)
commute]
= (T 2 I) (0) [By (1)]
= 0
and (T 1 I) (T 2 I) (v2) = (T 1 I) ( 21 v1) [By (1)]
= 21 (T 1 I) (v1) = 0 [By (2)]
Hence , we have proved
(T 1 I) (T 2 I) (vk) = 0 , k = 1 , 2. ……..(3)
Again we compute
(T 1 I) (T 2 I) (T 3 I) (v1) = (T 2 I) (T 3 I) (T 1 I) (v1)
= (T 2 I) (T 3 I) (0) = 0
and (T 1 I) (T 2 I) (T 3 I) (v2) = (T 3 I) (T 1 I) (T 2 I) (v2)
= (T 3 I) (0) = 0 [By (3)]
and (T 1 I) (T 2 I) (T 3 I) (v3) = (T 1 I) (T 2 I) ( 31 v1 + 32v2) [By (1)]
= 31 (T 1 I) (T 2 I) (v1) + 32 (T 1 I) (T 2 I) (v2)
= 31.0 + 32 .0 = 0 [By (3)]
Continuing this type of calculations, we can obtain
(T 1 I) (T 2 I)………(T n I) (vk) = 0 for k = 1 , 2 ,………, n
Let S = (T 1 I) (T 2 I)………(T n I) , then we have
S (v1) = S(v2) =……….= S(vn) = 0
Now, S annihilates a basis of V, so S must annihilate all of V as follows :
Let v V be any element , then v = 1v1 + 2v2+…………+ nvn , i F, and we have
S(v) = 1 S(v1) + 2 S(v2)+………..+ n S(vn) = 0
S(v) = 0 v V S=0
(T 1 I) (T 2 1 I)................(T n I) = 0 ……..(4)
Consider the polynomial
f(x) = (x 1) (x 2)………(x n).
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 17
Then, clearly f(x) F [x] and deg f(x) = n, and also T satisfies f(x), which is clear by
(4).
Def. Nilpotent Transformation : A linear transformation T A(V) is said to be nilpotent
transformation if there exist a positive integer k such that Tk = 0.
Theorem 1 : If S and T are nilpotent linear transformations such that S T = T S, then S T
and
S + T are nilpotent linear transformations.
Proof : Since S and T are nilpotent, so there exist positive integers n and m such that
Sn = 0 and Tm = 0 ……..(1)
(i) To prove that S T is nilpotent, consider
(ST)m = ST ST……… ST = (S.S……….S) (TT………T) [ ST = TS]
m m m
= S . T = S .0=0 [By (1)]
which shows that S T is nilpotent.
(ii) Now, we shall prove that S + T is nilpotent. Since ST = TS , so for any k, we have
k
k --1 k --r k --r
(S + T)k = k C0 Sk k C1 S T+.......+ k Cr S Tr +......+ k Ck Tk = k
Cr S Tr
r 0
…..(2)
m n --1
We shall prove that (S+T) = 0. For this let us take k = m+n 1 in (2) to obtain
m n --1
m n --1 m n --1 m n --1--r
(S+T) = Cr S Tr
r 0
m 1 m n --1
m n --1 m n --1--r r m n --1 m n --1--r
= Cr S T + Cr S Tr ....(3)
r 0 r m
Now , by (1), Sn = 0 , Tm = 0 implies that
Sk = 0 for k n and Tk = 0 for k m.
m n --1--r
Now, if 0 r m 1 , then m 1 r 0 m+n 1 r n and so S = 0
….…(4)
Also, if r m then clearly Tr = 0 ……….(5)
Using values of (4) and (5) , in (3), we obtain
m n --1
(S+T) = 0+0 = 0, which shows that S +T is nilpotent.
Remark : In first part of above theorem, it is very much clear that nilpotency of any one of
S and T is sufficient for S T to be nilpotent.
Now, we give some lemmas to prove an important theorem.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 18
Lemma 1 : If V = V1 V2 ………. Vk , where each subspace Vi is of dimension ni
and is invariant under T A(V), then a basis of V can be found so that the matrix of
T in this basis is of the form
A 0 0
1
0 A2 ........0
........................
0 0 Ak
where each Ai is a ni ni matrix and is the matrix of the linear transformation induced byT
on Vi.
Proof : We shall prove the lemma by induction on k. Suppose k = 2 and let
V = V1 V2 , where dim V1 = n1 , dim V2 = n2 and V1 , V2 are invariant
under T. Let u , u ,., u
1 2 n1
and v , v ,., v be basis of V and V over F
1 2 n2 1 2
respectively, Then, since V = V1 V , so u , u ,., u , v , v ,., v is a basis of V
2 1 2 n1 1 2 n2
over F. We claim that this is our required basis.
Now, we are given that V1 , V2 are invariant under T, so
T(ui) V1 , 1 i n1 and T(vj) V2 , 1 j n2 .
So , we can write
T u1 11u1 12u2 ...... 1n un 0.v1 0.v2 ...... [Link]
1 1 2
T u2 21u1 22u2 ... 2n un 0.v1 0.v2 ... [Link]
1 1 2
………………………………………
T un n ,1 u1 n ,2 u2 .... n ,n un 0.v1 0.v2 .... 0 .vn
1 1 1 1 1 1 2
T v1 0.u1 0.u2 ...... [Link] 11v1 12v2 ...... 1,n vn
1 2 2
T v2 0.u1 0.u2 ... [Link] 21v1 22v2 ... 2,n vn
1 2 2
…………………………………….
T vn 0.u1 0.u2 . [Link] n ,1v1 n ,2v2 . n
2 1 2 2 2 ,n2
vn
2
Let us suppose A = ij and B = b
n1n1 ij n2n2
Then, by definition of matrix of linear transformation, the matrix of T on V in the basis
A O
u1, u2 ,........., un , v1 ,v2,.........., vn
1 2
is
O B
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 19
where clearly A is matrix of T on V1 in the basis u1, u2 ,........, un
1
and B is matrix of
T on V2 in the basis v1 ,v2,..........., vn
2
. So, the theorem is true for k = 2.
Now let us assume as our induction hypothesis that theorem is true for k –1.
Now, let V V1 V2 .... Vk = V1 W where W V2 V3 ..... Vk
We have, V2 ,V3 ,..........., Vk are invariant under T, so their direct sum i.e. W is also
invariant under T. Therefore by induction hypothesis, there exist a basis of W in
which matrix of T is of the form
A O
2 O ...
O A ... O
3
... ... ... ...
O O ... Ak
where Ai is matrix of T on Vi of order ni ni , 2 i k.
Hence by the result proved for k = 2, we can obtain a basis of V in which the matrix of
T is of the form
A O ... O
1
O A ... O
2
... ... ... ...
O O ... Ak
where each Ai is an ni ni matrix and is matrix of T on Vi.
Lemma 2 : If T A(V) is nilpotent and 0 0, then 0 I + 1 T +..........+ m Tm is
invertible.
Proof : If 0 0 and S A(V) be any element s.t. Sr = 0, then we notice that
I S S2 r --1 S
r --1
( 0 I + S) 2 3 .........+(--1) r
0
0
0 0
r --1 r
S S S2 S2 S3 S3 r --1 S r --1 S
= I .........+ ( – 1) ( – 1) = I
0 0 02 02 03 03 0
r --1
0r
0 I + S is invertible.
Hence we proved,if 0 0 and S A(V) is s.t. Sr = 0,then 0 I + S is invertible .....(1)
Now, T is given to be nilpotent , so there exist some positive integer, say r, s.t. Tr = 0.
Let S = 1 T + 2 T2 .........+ m Tm , then we have
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 20
r
Sr = 1
T + 2 T2 ..........+ m Tm = 0
[In every term of expansion, the power of T is r, so every term becomes 0].
Sr = 0
So by (1), 0 I + S must be invertible which implies that
0 I + 1 T + 2 T2 ..........+ m Tm is invertible. This completes the proof.
Def. Index of nilpotency : If T A(V) is nilpotent, then k is called the index of nilpotency
of T if Tk = 0 but Tk–1 0.
Lemma 3 : If T A V is nilpotent of index n1 , then there exists a vector v V such that
n11
v,T (v ), T2 (v ),........,T (v ) are L. I.
n1 n11 n11
Proof : Since T 0 but T 0 , we can find a vector v V s.t. T (v ) 0 .
We claim that v,T (v ), T2 (v ),........,T n11(v ) are linearly independent. Let, if possible, these
are linearly dependent, then there exist scalars 1, 2,.........., n , not all zero, such that
1
n1 1
1v 2T (v ) ........ n T (v ) 0
1
Let s be the first non-zero scalar, then above expression reduces to
s --1 n1 1
s T (v ) s 1 Ts (v ) ........ n T (v ) 0
1
s --1 n s
(v ) s I s 1T ......... n T 1 0
T .........(1)
1
Since s 0 and T is nilpotent , so by lemma (2),
n1 s
s I s 1 T .......... n T is invertible and so must be non-zero.
1
Hence, by (1) , we have T s 1(v ) 0 . But s – 1 n1 which is a contradiction to the
fact that index of nilpotency of T is n1 . Hence, v,T (v ), T2 (v ),........,T n11(v ) are L.I.
Lemma 4 : If T A V is nilpotent of index of nilpotencey n1 and let v V be such that
n11
v,T (v ), T2 (v ),........,T (v ) are linearly independent (as in lemma (3)), and suppose V1 is
a subspace of V generated by these elements. If u V1 is such that T
n1k
u 0 where
0 k n1 , then there exist a vector u 0 V1 s.t. u T k u0 .
Proof : Since u V , so we can write
k --1 n1 1
u 1v + 2T(v ) ...........+k T (v )+k 1Tk (v ) ........... n T (v) ......... (1)
1
and then
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 21
T
n1 k
u 1Tn k (v) 2Tn k 1(v) ..... k Tn 1(v) k 1Tn (v) ...... n T2n k 1(v)
1 1 1 1
1
1
n1 k
But T r 0 if r n1 and T (u ) 0 , so we obtain, by above expression, that
n1 k n1 k 1 n1 1
0 1T (v ) 2T (v ) ....... k T (v ) .........(2)
But T n1 k
(v ),......,T
n1 1
(v ) , being a subset of linearly independent set
v,T (v),....,T n1 1
(v ) must be linearly independent. So, all scalars in expression (2) must
vanish i.e.
1 2 .......... k 0
Using these in (1) , we obtain
n1 1
u k 1 T k (v ) k 2 T k 1(v ) ........ n T (v )
1
T k k 1.v k 2 T (v ) ........ n T
1
n1 k 1
(v )
k
T (u0 ) , say , where
u 0 k 1.v k 2 T v ...... n T v V .
n1 k 1
1
1
Hence u T u 0 , where u 0 V1 .
Remark : V1 is invariant under T.
Proof : Let v1 V1 be any element, then
v1 1v 2 T v ......... n T v
n1 1
1
T v1 1T v 2T 2 v .......... n 1 T v v
n1 1 n1
and so n1
T
1
T 1 0 , so T v1 1T v 2T 2 v .......... n 1 T n1 1 v V1
n
But we have ,
1
T v , T v ,,T v are generators of
2 n11
because V1. Hence V1 is invariant under
T.
Lemma 5 : There exist a subspace W of V which is invariant under T and is such that
V V1 W , where T and V1 are same as defined in above lemmas.
Proof : Let W be a subspace of V of largest possible dimension such that
(i) V1 W 0
(ii) W is invariant under T.
To prove the lemma it is sufficient to prove that V = V1 + W. Let if possible, V V1
+ W, then there exist an element z V s.t. z V1 + W. Now given that Tn 1 z 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 22
V1 + W and T0(z) = z V1 + W and therefore there exist a smallest positive integer k
such that
Tk(z) V1 + W and Ti(z) V1 + W for i < k .........(1)
Thus, Tk(z) = u + w for some u V1 , w W .........(2)
T
n1 k
T (z )
k
= T
n1 k
(u) + T
n1 k
(w)
n1 k n1 k
0= T (u) + T (w) …….(3)
n1 1 n1
Since V1 is generated by v , T(v) , …….., T (v) and T = 0 so V1 is invariant under
n1 k
T, and therefore T (u) V1 ……..(4)
n1 k n1 k
Also, W is invariant under T, so T (w) W T (u) W [By (3)]
n1 k
T (u) W …….(5)
n1 k n1 k
By (4) and (5), we have, (u) V1 W = {0}
T T (u) = 0
So , by lemma (4), u = Tk(u0) for some u0 V1.
Then , by (2), Tk(z u0) = w Tk(z1 ) = w W , where z1 = z u0.
Now ,W is invariant under T, so for m k, Tm(z1) = T
m --k
T (z ) = T
k
1
m --k
(w) W
On the other hand, if i < k, then Ti (z1) = Ti(z) Ti (u0).
Now, V1 is invariant under T, so Ti(u0) V1 V1 + W and if Ti (z1) V1 + W ,
then we must have Ti (u0) + Ti (z1) = Ti (z) V1 + W , which is contradiction to
choice of k .
Thus , we have computed that
Tm (z1) W V1 + W for m k and Ti (z1) V1 + W for i < k ........(6)
k–1
Let W1 be the subspace of V spanned by W and z1 , T (z1) ,…….., T (z1).
Since z1 W , so W
W1 and therefore dim (W1) > dim (W).
Now, we prove that W1 is also invariant under T. Let w W1 be any element, then we
write w = w0 + 1z1 + 2 T(z1) +...........+ k Tk–1 (z1) , where w0 W
and so T (w) = T (w0) + 1 T(z1) + 2 T2 (z1) +………+ k Tk(z1).
Now, W is invariant under T, so T(w0) W W1.
Further T (z1) , T2(z1) ,………,Tk–1 (z1) are generators of W1 and also Tk(z1) W W1
.…(6)
Thus, T(w) W1 showing that W1 is invariant under T. Now dim (W1) > dim (W)
and W1 is also invariant under T, so, if V1 W1 = {0} , then maximal nature of
W is contradicted. So, we have V1 W1 {0}
Let w ( 0) V1 W1 , then w W1 and so it must be of the form
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 23
w = w0 + 1z1 + 2 T(z1) +.............+ k Tk–1 (z1), where w0 W. ...........(7)
Not all of 1, 2 ,............, k can be 0 : otherwise we would have
0 w = w0 W V1 = {0} [ w0 W] , a
contradiction.
So, let s be the first non – zero , then by (7),
s --1 k --1
w 0 s T (z 1 ) s 1 Ts (z 1) ........ k T (z 1 ) V1
k --s s --1
w0 + s I s 1 T ......... k T T (z1) V1 ……..(8)
Since s 0 , by lemma-(2), we have
k --s
s I s 1 T ......... k T is invertible. Let its inverse by R and R will be
a polynomial in T.
Now, W and V1 are invaunder T, so W and V1 are invariant under R also.
Applying, R on both sides of (8), we get , R(w0) + Ts–1 (z1) V1
Since W is invariant under T, so, R(w0) W
So we have Ts–1(z1) V1 + W Ts–1 (z u0) V1 + W [ z1 =
z u0 ]
Ts–1(z) Ts–1(u0) V1 + W
But Ts–1(u0) V1 V1 + W and so finally we have Ts–1(z) V1 + W
which is a contradiction by equation-(1), because s 1 < k.
Therefore, we must have V = V1 + W. Also V1 W {0} and so V = V1 W.
0 1 0 ... 0 0
0 0 1 ... 0 0
Notation : Mt will denote the matrix ... ... ... ... ... ...
0 0 0 ... 0 1
0 0 0 0 0 0
all of whose entires are 0 except on the super diagonal, they are 1’s.
Theorem 1 : If T A(V) is nilpotent, of index of nilpoency n1 , then a basis of V can be
M 0 ... 0
n1
0 Mn ... 0
found such that the matrix of T in this basis has the form 2
... ... ... ...
0 0 ... Mn
r
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 24
where n1 n2 ........ nr and n1 n2 nr = dim V.
n n1 1 n1 1
Proof : Since T 1 = 0 and T 0, we can find a vector v V such that T (v)
n1 1
0. Then by Lemma 2, v , T(v),…….., T (v) are linearly independent. Let V1 be the
n1 1
vector subspace of V generated by v, T(v) , T2(v) , ………, T (v).
n1 1
Let us denote v1 = v , v2 = T(v) ,………, vn = T (v).
1
Then V1 is invariant under T and we can write
T(v1) = 0. v1 + 1. v2 + 0. v3 +………+ 0. vn–1 + 0. vn
T(v2) = 0. v1 + 0. v2 + 1. v3 +………..+ 0. vn–1 + 0. vn
……………………………………………………
T(vn–1) = 0. v1 + 0. v2 + 0. v3 +……….+ 0. vn–1 + 1. vn
T(vn) = 0. v1 + 0. v2 + 0. v3 +………..+ 0. vn–1 + 0. vn
So , the matrix of T on V1 in the basis {v1, v2,., vn } is
1
0 1 0 ... 0 0
0 0 1 ... 0 0
Mn ... ... ... ... ... ...
1
0 0 0 ... 0 1
0 0 0 ... 0 0
n1n1
Now by Lemma (5), V = V1 W where W is invariant under T. Using the basis {v1 , v2 ,
….., vn }
1
of V1 and any basis of W as a basis of V and applying Lemma (1) , the matrix of T in
M 0
n1
this basis has the form
0 A2
where A2 is the matrix of T2 on W where T2 is the linear transformation induced by T
n1 n
on W. Since T = 0 , so T2 2 = 0 for some positive integer n2 n1.
Repeating the argument used for T on V for T2 on W , we can decompose W as we
did V. Continuing this way , we get a basis of V in which the matrix of T is of the form
M 0 ... 0
n1
0 Mn ... 0
2
... ... ... ...
0 0 ... Mn
r
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 25
Also, n1 n2 .nr = dim V1 + dim V2 +………+ dimVr = dim V.
Def. Invariants : The integers n1, n2 ,..nr in above theorem-1 are called invariants of
T.
Def. Cyclic subspace : If T A(V) is nilpotent, the subspace M of V, of dimension m,
which is invariant under T , is called cyclic with respect to T if
(i) Tm (M) = {0} , Tm–1 (M) {0}
(ii) there is an element z M such that z , T(z) ,………,Tm–1(z) form a basis of M.
Remark : Theorem 1 tells us that given a nilpotent T A(V) we can find integers
n1 n2 ........ nr and subspaces V1,V2 , .,Vr of V cyclic with respect to T and of
dimensions n1, n2 ,..nr respectively such that V = V1 V2 ………. Vr
Lemma 6 : If M, of dimension m, is cyclic with respect to T, then the dimension of Tk
(M) is m–k for all k m.
Proof : Since M is cyclic w.r.t. T, so, by definition,
(i) Tm (M) = {0} , Tm–1 (M) {0}
(ii) there is an element z M such that z , T(z) , T2(z)………,Tm–1(z) form a basis of M.
Now, we can find a basis of Tk(M) by taking image of any basis of M under Tk(M).
Now, by (ii) , {z , T(z) , ……….,Tm–1(z)} is a basis of M and so
{Tk(z) , Tk+1(z) ,………., Tm–1(z) , Tm(z) ,………, Tm+k–1(z)} must be a basis of Tk(M).
But Tm(z) = Tm+1(z) = ……….. = Tm+k–1(z) = 0 [By (i)]
and so basis of Tk(M) is { Tk(z) , Tk+1 (z) ,……….. Tm–1 (z)} which contains m k elements.
Hence we get dim Tk(M) = m – k. This completes the proof.
Theorem 2 : Invariants of a nilpotent transformation are unique.
Proof : Let n1 n2 ........ nr and m1 m2 ........ ms be two sets of invariants of T.
Then there exist subspaces V1 , V2 , ………., Vr of V cyclic w.r.t. T and of dimensions
n1 n2 ........ nr respectively such that V = V1 V2 ……… Vr
Similarly, there exist subspaces U1 , U2 , ………., Us of V cyclic with respect to T and of
dimensions m1 m2 .......... ms respectively such that V = U1 U2 ……….. Us
We shall prove that r = s and m1 = n1 , m2 = n2 ,………., mr nr .
Let, if possible, this is not true, then there is a first integer i such that
m1 = n1 , m2 = n2 ,..........., mi1 = ni–1 , mi ni . .........(1)
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 26
W.L.O.G. , suppose m i ni .
Now, we compute T
mi
V in two different ways.
We have V = V1 V2 ……… Vr
m mi mi mi
T i (V ) = T V1 T (V2 ) ………… T (Vr)
mi mi mi mi
dim T (V) = dim T (V1 )+ dim T (V2 )+……….. + dim T (Vr)
mi mi mi
dim T (V1 )+ dim T (V2 )+………..+ dim T (Vi)
Making use of Lemma (6) , we obtain
mi
dim T (V) (n1 m ) + (n2 mi ) +..............+( ni mi ) ..........(2)
i
On the other hand, V = U1 U2 ................. Us and so
mi mi mi mi
T (V) = T (U1) T (U2) .............. T (Us)
mi m mi
= T(U1) T i (U2) ................ T (Ui–1)
Tmi (U ) Tmi (U ) = ...... = 0
i i 1
mi mi mi mi
and therefore, dim T (V)=dim T (U1 )+ dim T (U2 )+ ........... +dim T (Ui–1)
Again, using Lemma (6) , on R.H.S. , we get
mi
dim T (V) = (m1 mi ) + (m2 mi ) + ......... + ( mi 1 mi )
= (n1 mi ) + (n2 mi ) + ......... + (ni–1 mi ) [By (1)]
Using this in (2), we obtain 0 ni mi , which is a contradiction since ni mi > 0.
Hence invariants of a nilpotent transformation are unique.
Def. Similar Linear Transformations : The linear transformations S, T A(V) are said to be
similar if ther exist an invertible element C A(V) s.t. T = C S C–1.
Def. Similar matrices : Two matrices A and B are said to be similar if there exist an
invertible matrix C s.t. B = C A C1 .
Results :
1. Let S and T be linear transformations on V such that the matrix of S in one basis
is equal to the matrix of T in other basis. Then S and T are similar.
2. Two nilpotent transformations are similar iff their matrices in some fixed basis are
similar.
3. If n1 n2 ........ nr and m1 m2 .......... ms are such that
n1 n2 . nr = m1 m2 ...... ms then the matrices
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 27
M 0 ... 0 M 0 ... 0
n1 n1
0 Mn ... 0 0 Mn ... 0
2
and 2
... ... ... ... ... ... ... ...
0 0 ... Mn 0 0 ... Mn
r s
are similar iff r = s and n1 = m1, ........... , nr = mr.
Theorem 3 : Two nilpotent linear transformations are similar iff they have the same
invariants.
Proof : Let S , T A(V) and dim V = n.
Suppose first that S and T are similar. We shall prove that they have the same
invariants.
Let n1 n2 ........ nr and m1 m2 .......... ms be invariants of S and T
respectively. Then by theorem-(1) , matrices of S and T in one basis will have the form
M 0 ... 0 M 0 ... 0
n1 n1
0 Mn ... 0 0 Mn ... 0
2
and 2
... ... ... ... ... ... ... ...
0 0 ... Mn 0 0 ... Mn
r s
Since S and T are similar, so their respective matrices must also be similar. Then, we
must have r = s and n1 = m1 , n2 = m2 , ........... , nr = mr.
Hence invariants of S and T are same.
Conversely : Suppose two nilpotent transformations S and T have same invariants
n1 n2 ........ nr . Then by theorem-(1) , there are bases {v1 , v2 , ......... , vn}
and {w1 , w2 , .........., wn} of V such that matrix of S in {v1 , v2 , ........ vn} and that of
M 0 ... 0
n1
0 Mn ... 0
T in {w1 , w2 , .......... , wn} are each equal to 2
... ... ... ...
0 0 ... Mn
s
But we know that if matrices of two linear transformations in any basis are same, then two
linear transformations are similar linear transformations.
Remark : Let Fn denote the set of all n n matrices over the field F. Two matrices A and
B are said to be similar if there exists an invertible matrix C Fn s.t. B = CA C1 .
Further this relation of similarity is an equivalence relation and similarity class of a matrix A
is the set of all matrices similar to A.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 28
Def. Partition of an integer : Let n be a positive integer. A sequence of positive integers
n1, n2,......, nk ; n1 n2 ....... nk , such that n n1 n2 ........ nk is called a partition of
n and n1, n2,......, nk are called parts of the partition.
For example, let n = 3. Then 3 = 1 + 1 + 1 , 3 = 1 + 2 , 3 = 3 are all partitions of n = 3.
This gives three partitions of n. Also n = 4 has 5 partitions namely,
4 = 1 + 1 + 1 + 1 , 4 = 1 + 1 + 2 , 4 = 1 + 3, 4 = 2+2, 4 = 4. The number of partitions
of n is denoted by p(n). So p(3) = 3 , p(4) = 5.
Theorem 4 : Number of distinct similarity classes of nilpotent n n matrices is precisely p(n)
, the number of partitions of n.
Proof : Clearly, it is sufficient to prove the theorem for linear transformations. In fact we
shall prove that number of distinct similarity classes of nilpotent linear transformations on a
n-dimensional vector space is precisely p(n), the number of partitions of n.
Let T A(V), where V is of dimension n and T is nilpotent of index of nilpotency n1 . We
have proved that if n1 n2 ....... nr are invariants of T, then n = n1 n2 ....... nr i.e.
invariants of T form a partition of n. Conversely, any partition of n , n1 n2 ....... nr s.t.
n1 n2 ....... nr = n determines the invariants of nilpotent linear transformations given
M 0 ... 0
n1
0 Mn ... 0
by the matrix 2
... ... ... ...
0 0 ... Mn
r
So we can say that every nilpotent linear transformation gives rise to a partition of n and
every partition of n gives rise to a nilpotent matrix (linear transformation). Also two
nilpotent linear transformations are similar if and only if they have same invariants. So
number of distinct similarity classes of nilpotent n n matrices is precisely p(n) , the number
of partitions of n.
Primary Decomposition Theorem and Jordan form
Lemma 1 : Let T A(V) and V1 is a subspace of V invariant under T. Let T induces the
linear transformation T 1 on V1 . Then prove that,
(i) for any polynomial q(x ) F( x ) , the linear transformation induced by q(T) on V1 is
precisely q(T 1 )
(ii) the minimal polynomial of T 1 divides minimal polynomial of T.
Proof : (i) Since T 1 is the linear transformation induced by T on V1 , so we must have,
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 29
T (v) = T 1 (v) for all v V1
Now, T2 (v) = T( T(v) ) = T( T 1 (v) ) = T 1 ( T 1 (v) ) [ T 1 (v)
V1 ]
= T12 (v)
Continuing like this, we get, Tn (v) = T1n (v) for all v V1
Now for any polynomial, q(x ) = 0 1x ............. nx n and for any v V1 , we see
that,
q(T) (v) = ( 0 I 1T ............. n Tn ) (v)
= 0 I(v ) 1T(v ) ........ n Tn (v )
= 0 I(v ) 1T1(v ) ........ n T1n (v )
= 0I 1T1 ............. n T1n )(v) = q(T 1 ) (v)
Thus q(T) (v) = q(T 1 ) (v) for all v V1 and so q(T 1 ) is the linear transformation induced
by q(T) on V1 .
(ii) Let m(x ) and m1(x ) be the minimal polynomials of T and T 1 respectively. Then we
must have, m(T) = 0 and m1(T1 ) = 0
Now for any v V1 , we see that , m(T 1 ) (v) = m(T) (v) [by part
(i)]
= O (v) = O
m(T 1 ) = O T 1 satisfies m(x ) . But m1(x ) is the minimal polynomial of T 1 , so
m1(x ) m(x ) .
Lemma 2 : Suppose that V = V1 V2 , where V1 and V2 are subspaces of V invariant
under T. Let T 1 and T 2 be the linear transformations induced by T on V1 and V2
respectively. If the minimal polynomial of T 1 over F is p1(x ) while that of T 2 is p2 (x ) , then
the minimal polynomial for T over F is the least common multiple of p1(x ) and p2 (x ) .
Proof : Let q(x ) be the l.c.m. of p1(x ) and p2 (x ) and if p(x ) be the minimal polynomial of
T over F , then by above lemma, p1(x ) p(x ) and p2(x ) p(x )
l.c.m. p1(x ), p2 (x ) p(x ) q(x ) p(x )
………(1)
Again, let v V be an arbitrary element , then v = v1 + v2 , v1 V1 , v2 V2
q(T) (v) = q(T) ( v1 ) + q(T) ( v2 ) = q(T 1 ) ( v1 ) + q(T 2 ) ( v2 ) ………(2)
Now, since q(x ) = l.c.m. p1(x ), p2 (x ) , so we have, p1(x ) q (x ) and p2(x ) q (x )
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 30
But p1(T 1 ) = 0 and p2 (T 2 ) = 0 , so q(T 1 ) = 0 and q(T 2 ) = 0 .
Using this in (2), q(T) (v) = 0 + 0 = 0
Hence we have proved, q(T) (v) = 0 for all v V
q(T) = 0 T satisfies q(x ) .
But p(x ) is the minimal polynomial of T , so p(x ) q (x ) ………(3)
By (1) and (3) , p(x ) = q(x ) [ p(x ) is monic]
Hence minimal polynomial of T is the l.c.m. of minimal polynomials of T 1 and T 2 .
Corollary : If V = V1 V2 ............. Vk where each Vi is invariant under T and if pi (x )
is the minimal polynomial over F of T i , the linear transformation induced by T on Vi , then
the minimal polynomial of T over F is the least common multiple of p1(x ) , p2 (x ) , ……. ,
pk (x ) .
Proof : We have already proved the result for n = 2. By induction, corollary is proved.
Theorem 1 : Primary Decomposition Theorem : Let T A F (V) and suppose p(x ) F[ x ] be
the minimal polynomial for T such that p x q1(x ) 1 q2(x ) 2 .................. qk (x ) k , where the
l l l
qi (x ) are distinct irreducible polynomials in F[ x ] and li are positive integers.
l
Let Vi be the kernels (or null spaces) of qi (T) i , i = 1, 2, …… , k . Then
(i) each Vi is invariant under T
(ii) V = V1 V2 ............. Vk
(iii) if Ti be the linear transformation induced byT on Vi , then the minimal polynomial of
l
Ti is qi (x ) i .
Proof :
(i) We have, Vi = v V : qi (T) i (v ) 0
l
l l l
Let v Vi be an element, then , qi (T) i (Tv) = T qi (T) i (v ) [since Ti and qi (T) i
commute]
= T (0) = 0
T (v) Vi Vi is invariant under T.
(ii) If k = 1 , then V = V1 and there is nothing to prove.
So let k > 1. We define k polynomials,
l l l
h 1(x ) = q2(x ) 2 q 3(x ) 3 ....... qk (x ) k
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 31
l l l
h 2(x ) = q1(x ) 1 q 3(x ) 3 ........ qk (x ) k
. …………………
lj
h i (x ) = q j (x )
j i
………………..
l l l
h k (x ) = q1(x ) 1 q2(x ) 2 ........qk 1(x ) k 1
First we prove that each Vi {0}. Since k > 1, so deg h i (x ) < deg p(x ) and hence
h i (T) 0.
Thus there must be an element v V such that h i (T) (v) 0 and let h i (T) (v) = w
l l
But then, qi (T) i (w) = qi (T) i h i (T) (v) = p (T) (v) = O(v) = 0
w ( 0) Vi Vi {0}.
Now we claim that, h i (T) (V) Vi ………(1)
For this, let h i (T) (v) , v V be an arbitrary element. If h i (T) (v) = 0 , then obviously 0
Vi and if h i (T) (v) 0 , then by above computations, h i (T) (v) Vi . Hence the claim.
l l
Now for j i , we see that q j (x ) j hi (x ) and Vi = v V : q j (T ) j (v ) 0
so hi (T )(v j ) = 0 for all v j V j ……..(2)
Now we claim that, V = V1 V2 ....... Vk .
For this, we see that qi (x ) are irreducible elements and one of these is absent in each h i (x ) ,
so h 1(x ) , h 2(x ) , ……, h k (x ) are relatively prime. There must exist polynomials
a1(x ), a2(x ),......, ak (x ) in F[ x ] such that a1(x )h1(x ) a2(x )h2(x ) ....... ak (x )hk (x ) 1
From this we get, a1(T)h1(T) a2(T)h2(T) ........ ak (T)hk (T) I
Now let v V be an arbitrary element, then
v = I (v) = a1(T)h1(T) a2 (T)h2 (T) ........ ak (T)hk (T) (v)
= h1(T)a1(T) h2(T)a2 (T) ......... hk (T)ak (T) (v)
= h1(T)a1(T)(v ) h2(T) a2(T)(v ) ....... hk (T) ak (T)(v )
V1 V2 ........ Vk [By (1)]
V V1 V2 ....... Vk
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 32
Combining this with V1 V2 ........ Vk V , we obtain, V = V1 V2 ........ Vk
We must now verify that this sum is direct sum. To show this, it is sufficient to prove that if
u1 u2 ........ uk 0 with each ui Vi , then each ui 0 .
Let , if possible, some ui , say u1 0, then
h 1(T) ( u1 u2 ....... uk ) = h 1(T) (0)
h1(T)(u1 ) h1(T)(u2 ) ....... h1(T)(uk ) = 0
h1(T)(u1 ) = 0 [By (2)]
l l
Also, q1(T) 1 (u1) = 0. But since h 1(x ) and q 1(x ) are relatively prime , so h 1(x ) and q1(x ) 1
must also be relatively prime and hence there must exist two polynomials a(x ) and b(x ) in
F[ x ] s.t.
l
a(x ) h 1(x ) + b(x ) q1(x ) 1 = 1
l
a(T) h 1(T) + b(T) q1(T) 1 = I
l
( a(T) h 1(T) + b(T) q1(T) 1 ) (u1 ) = I (u1 )
l
a(T) h 1(T) (u1 ) + b(T) q1(T) 1 (u1 ) = u1
0 + 0 = u1 u1 = 0, a contradiction.
Hence V = V1 V2 ............. Vk .
l
(iii) By the definition of Vi , qi (T) i (v) = 0 for all v Vi
l
qi (Ti ) i (v) = 0 for all v Vi [by lemma (1)]
l
qi (Ti ) i = 0
l l
Ti satisfies qi (x ) i and so minimal polynomial of Ti , say mi (x ) , must divide qi (x ) i .
r
mi (x ) = qi (x ) i , where ri li
Now we know that minimal polynomial of T is l.c.m. of minimal polynomials of Ti and so
r
p(x ) = l.c.m. q1(x ) 1 , q2(x ) 2 , ........,qk (x ) k
r r
r r r
p(x ) = q1(x ) 1 q2 (x ) 2 ......... qk (x ) k
[since qi ' s are irreducible and hence relatively
prime]
l l l r r r
q1(x ) 1 q2(x ) 2 ......... qk (x ) k = q1(x ) 1 q2 (x ) 2 ......... qk (x ) k
l1 r 1 , l2 r 2 , ……. , lk r k [since qi' s are irreducible]
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 33
l l
mi (x ) = qi (x ) i i.e. minimal polynomial of Ti is qi (x ) i .
Result : If V is n – dimensional over F and if T A(V) has all its characteristic
roots in F, then T satisfies a polynomial of degree n over F, namely
f(x) = (x 1) (x 2)………(x n). By this, it is clear that If all the characteristic roots of
T lie in F, then minimal polynomial of T takes the form,
l l l
p(x ) = (x 1) 1 (x 2 ) 2 ......... (x k ) k where 1, 2,......, k are the distinct
characteristic roots of T.
Corollary : If all the distinct characteristic roots 1, 2,......, k of T lie in F, then V can be
li
expressed as V = V1 V2 ........ Vk , where Vi =
v V : T i (v ) 0 and Ti has
only one characteristic root i on Vi .
Proof : If all the characteristic roots of T lie in F, then we know that minimal polynomial of
T takes the form,
p x (x 1) 1 (x 2 ) 2 ........ (x k ) k
l l l
where 1, 2,......, k are the distinct characteristic roots of T. By replacing the irreducible
factors qi (x ) by x i in primary decomposition theorem, we obtain the required result. Also
by construction of Vi itself, clearly, Ti has only one characteristic root i on Vi .
Def. Basic Jordan Block : The matrix ... ... ... ... ... with ' s on the diagonal, 1’s on
... ... ... ... 1
0 ... ... ...
the superdiagonal, and 0’s elsewhere, is a basic Jordan block belonging to .
Def. Jordan Block : A m m basic Jordan block belonging to is clearly I + Mm where
Mm is a m m matrix having superdiagonal elements 1 and 0 elsewhere.
Def. Jordan form : A matrix A is said to be in Jordan form if it can be represented as
J B
1 i1
J B
A = 2
where each J i i2
Jk Bir
i
and where Bi1, Bi 2,......... ,Bir are basic Jordan blocks belonging to i and in each Ji the size
i
of Bi 1 size of Bi 2 ……….
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 34
Theorem 2 : Let T A F (V) has all its distinct characteristic roots 1, 2,......, k in F. Then
a basis of V can be found in which the matrix T is of the form
J B
1 i1
J B
2 where each J = i2
i
Jk Bir
i
and where Bi 1, Bi 2,...... ,Bir are basic Jordan blocks belonging to i .
i
OR
Let T A F (V) has all its distinct characteristic roots 1, 2,......., k in F. Then a basis of V
can be found in which the matrix of T is in Jordan form.
Proof : We know by corollary to Primary decomposition theorem that, “If all the distinct
characteristic roots 1, 2,..., k of T lie in F, then V can be expressed as V= V1 V2 .... Vk ,
l
where Vi v V : T i i (v) 0 and Ti has only one characteristic root i on Vi .”
Then, clearly, Ti i is nilpotent on Vi and so using the theorem, “ If T A(V) is
nilpotent, of index of nilpoency n1 , then a basis of V can be found such that the matrix of
M 0 ... 0
n1
0 Mn ... 0
T in this basis has the form 2
... ... ... ...
0 0 ... Mn
r
where n1 n2 ...... nr and n1 n2 ...... nr = dim V ” , there is a basis of Vi in
M
i1
M
which matrix of Ti i is of the form i2
Mir
i
Further , Ti = i + ( Ti i ) and so matrix of Ti is of the form
M B
i i1 i1
M B
i + i2 = i2 = J
i
i Mir Bir
i i
Now using the lemma, “ If V = V1 V2 ….. Vk , where each subspace Vi is of
dimension ni and is invariant under T A(V), then a basis of V can be found so that
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 35
A 0
1 0 ...
0 A ... 0
the matrix of T in this basis is of the form 2
where each Ai is a ni
... ... ... ...
0 0 ... Ak
ni matrix and is the matrix of the linear transformation induced by T on Vi.” we can
say that a basis of V can be found in which the matrix of T is of the form
J B
1 i1
J B
2 where each J = i2
i
Jk Bir
i
and where Bi1, Bi 2,......... ,Bir are basic Jordan blocks belonging to i .
i
Clearly in each Ji , basic Jordan block can be arranged in order of non increasing size and
hence matrix of T is in Jordan form.
Remark : For nilpotent transformations, above theorem reduces to Theorem 1 of last topic.
1 1 1
Example 1 : Prove that the matrix 1 1 1 is nilpotent , and find its invariants
1 1 0
and Jordan form.
1 1 1
Solution : Let A = 1 1 1
1 1 0
1 1 1 1 1 1 1 1 0
A2 = A.A = 1 1 1 1 1 1 = 1 1 0
1 1 0 1
1 0
0 0 0
1 1 0 1 1 1 0 0 0
A3 = A2 .A = 1 1 0 1 1 1 = 0 0 0
0 0 0 1
1 0
0 0 0
Hence A is a nilpotent matrix with index of nilpotent 3.
For a n n nilpotent matrix of index of nilpitency n1 , we know that invariants are of the
form
n1 n2 ...... nr s.t. n1 n2 ...... nr = n and then matrix A is similar to
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 36
M 0 ... 0
n1
0 Mn ... 0
2
... ... ... ...
0 0 ... Mn
r
Now for the given matrix A we see that n = n1 = 3 , so A has only one invariant namely ,
0 1 0
3 and then A is similar to M3 i.e. 0 0 1 which is clearly a basic Jordan block
0 0 0
belonging to 0. Hence Jordan form of A is M3.
1 1 1
Example 2 : Prove that the matrix 1 1 1 is nilpotent and find its invariants
0 0 0
and Jordan form.
1 1 1
Solution : Let A= 1 1 1
0 0 0
1 1 1 1 1 1 0 0 0
A2 = A .A = 1 1 1 1
1 1 = 0 0 0
0 0 0
0 0 0
0 0 0
Hence A is a nilpotent matrix with index of nilpotency 2.
For a n n nilpotent matrix of index of nilpitency n1 , we know that invariants are of the
form
n1 n2 ...... nr s.t. n1 n2 ...... nr = n and then matrix A is similar to
M 0 ... 0
n1
0 Mn ... 0
2
... ... ... ...
0 0 ... Mn
r
Now for the given matrix A we see that n = n1 = 2 , so A has only one possible
arrangement of invariants, namely , n = n1 n2 = 2+1 and then A is similar to
0 1 0
M 0
2 = 0 0 0 which is clearly in Jordan form.
0 M1
0 0 0
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 37
Example 3 : Find all possible Jordan forms for all 8 8 matrices having x 2 (x 1)3 as
minimal polynomial.
Solution : Since distinct roots of characteristic polynomial and minimal polynimial are
always same , so characteristic polynomial can have four choices, namely,
x 5(x 1)3, x 4 (x 1)4, x 3(x 1)5, x 2(x 1)6
We consider these cases one by one,
(i) Charcterstic polynomial = x 5 (x 1)3 ; Minimal polynomial = x 2 (x 1)3
Here, at least one basic Jordan block belonging to ‘0’ must be of order 2 and sum of orders of
all basic Jordan blocks belonging to ‘0’ must be 5.
So either 5 = 2 + 2 +1 or 5 = 2+1+1+1.
Again, at least one basic Jordan block belonging to ‘1’ must be of order 3 and sum of orders
of all basic Jordan blocks belonging to ‘1’ must be 3. So 3 = 3 is the only possibility. So
two Jordan forms are possible in this case, namely,
1 1 0 1 1 0
0 1 0 1 0 1
diag.
,
,[0], 0 1 1 and diag.
,[0],[0],[0], 0 1 1
0 0 0 0 0 0 1 0 0 0 0 1
(ii) Charcterstic polynomial = x 4 (x 1)4 ; Minimal polynomial = x 2 (x 1)3
Here at least one basic Jordan block belonging to ‘0’ must be of order 2 and sum of orders of
all basic Jordan blocks belonging to ‘0’ must be 4. So either 4 = 2+2 or 4 = 2+1
+1 .
Again at least one basic Jordan block belonging to ‘1’ must be of order 3 and sum of orders
of all basic Jordan blocks belonging to ‘1’ must be 4. So 4 = 3 + 1 is the only possibility.
So two Jordan forms are possible in this case, namely,
1 1 0 1 1 0
0 1 0 1 , 0 1 1 ,[1] and diag. 0 1 ,[0],[0], 0 1 1 , [1]
diag. ,
0 0
0 0 0 0
0 0 1
0 0 1
(iii) Charcterstic polynomial = x 3 (x 1)5 ; Minimal polynomial = x 2 (x 1)3
Here at least one basic Jordan block belonging to ‘0’ must be of order 2 and sum of orders of
all basic Jordan blocks belonging to ‘0’ must be 3. So 3 = 2+1
Again at least one basic Jordan block belonging to ‘1’ must be of order 3 and sum of orders
of all basic Jordan blocks belonging to ‘1’ must be 5. So 5 = 3 + 2 or 5 = 3 + 1 + 1. So
two Jordan forms are possible in this case, namely,
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 38
1 1 0 1 1 0
0 1 0 1 1
0 1
0 1 1 ,[1],[1]
diag. ,[0], 1 1 , and diag. ,[0],
0 0 0 0 1 0 0
0 1
0 0 1
(iv) Charcterstic polynomial = x 2 (x 1)6 ; Minimal polynomial = x 2 (x 1)3
Here at least one basic Jordan block belonging to ‘0’ must be of order 2 and sum of orders of
all basic Jordan blocks belonging to ‘0’ must be 2. So 2 = 2.
Again at least one basic Jordan block belonging to ‘1’ must be of order 3 and sum of orders
of all basic Jordan blocks belonging to ‘1’ must be 6. So 6 = 3 + 3 or 6 = 3 + 2 + 1 or
6 = 3 + 1 + 1 + 1. So three Jordan forms are possible in this case, namely,
1 1 0 1 1 0
0 1 , 0 1 1
diag. , 0 1 1
0 0
0 0 1 0 0 1
1 1 0
0 1 , 1 1 ,[1]
diag. , 0 1 1
0 0 0 1
0 0 1
1 1 0
0 1 ,[1],[1],[1]
diag. , 0 1 1
0 0
0 0 1
1 1 1
Example 4 : Find the Jordan form of the matrix A = 1 1 1
1 0 0
Solution : Characteristic equation of A is given by A x I = 0
1 x 1 1
i.e. 1 1x 1 = 0
1 0 0 x
x (x –1) (x +1) = 0
So , characteistic polynomial is x (x –1) (x +1) and since all roots are distinct , so minimal
polynomial is same as the characteristic polynomial. Clearly , basic Jordan blocks
belonging to all three characteristic roots are of order 1 , so the Jordan form is
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 39
0 0 0
diag. [0] , [1] , [ 1] i.e. 0 1 0
0 0 1
Example 5 : Let V be a three-dimensional vector space over with
the complex field C
bassis {v1 , v2 , v3} and T L (V,V) be defined by :
T(v1) = –v1 + 2v3 , T(v2) = 3v1 + 2v2 + v3 , T(v3) = –v3
Find the Jordan cononical form of T.
1 0 2
Solution : Here , matrix of T in the basis {v1 , v2 , v3} is 3 2 1
0 0 1
Characteristic equation A is given by A x I = 0
1 x 0 2
i.e. 3 2x 1 = 0
0 0 1x
(x+1) (x+1) (x –2) = 0
So , characteristic polynomial of A is (x+1)2 (x –2).
Now ,minimal polynomial of A is either (x+1) (x –2) or (x+1)2 (x –2)
0 0 2 3 0 2 0 0 6
But (A+I) (A –2 I) = 3 3 1 3 0 1 3 0 6 0
0 0 0 0 0 3 0 0 0
2
so that minimal polynomial is (x+1) (x –2).
Now , clearly , basic Jordan block belonging to –1 is order of 2 and that of 2 is of order
1 1 0
1 1
1 , so Jordan form of A is diag = , [2] i.e. 0 1 0
0 1
0 0 2
1 1 1 1 1 1
Example 6 : Find the Jordan forms of 1 1 1 and 1 1 1 by using
1 1 0 0 0 0
characteristic and minimal polynomials.
Solution : Do yourself.
MATHS ZONE ACADEMY
Shop No. 58, Huda Market, Opposite DRDA, Near PNB Bank, Jind. Ph. 9896729467, 9896999278
[Link]. 2nd Sem. ALGEBRA UNIT-IV PAGE 40
Result : Two linear transformations in AF(V) which have all their characteristic roots in
F are similar if and only if they can be brought to the same Jordan form.
1 1 1 1 1 1
Example 7 : Prove that the matrix 1 1 1 is not similar to 1 1 1
1 1 0
1 0 0
Solution : Do yourself.
You might also like
- Mathematics Encyclopedia: Modules & Rings100% (1)Mathematics Encyclopedia: Modules & Rings1,290 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
- Understanding Modules and Vector SpacesNo ratings yetUnderstanding Modules and Vector Spaces75 pages
- Module Theory: Definitions and ExamplesNo ratings yetModule Theory: Definitions and Examples21 pages
- Understanding Rings and Modules in AlgebraNo ratings yetUnderstanding Rings and Modules in Algebra5 pages
- Understanding R-Modules and Their PropertiesNo ratings yetUnderstanding R-Modules and Their Properties21 pages
- Introduction to Modules Theory in AlgebraNo ratings yetIntroduction to Modules Theory in Algebra20 pages
- Unitary Zn-Modules and R-Module HomomorphismsNo ratings yetUnitary Zn-Modules and R-Module Homomorphisms4 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
- Essential Submodules and Their GeneralizationNo ratings yetEssential Submodules and Their Generalization13 pages
- Commutative Algebra: A-Modules ExplainedNo ratings yetCommutative Algebra: A-Modules Explained21 pages
- Aops Community 2008 China Team Selection TestNo ratings yetAops Community 2008 China Team Selection Test4 pages
- Matrix Norms and Transformations ExplainedNo ratings yetMatrix Norms and Transformations Explained17 pages
- Bipartite Graphs and Their ApplicationsNo ratings yetBipartite Graphs and Their Applications40 pages
- 3.6.3 Practice - Polynomial and Rational Functions (Practice)No ratings yet3.6.3 Practice - Polynomial and Rational Functions (Practice)8 pages
- Random Variables and Probability ConceptsNo ratings yetRandom Variables and Probability Concepts40 pages
- Equivalent Norms and Compactness in Finite DimensionsNo ratings yetEquivalent Norms and Compactness in Finite Dimensions10 pages