Module Collected

Chapter 3
Modules and Vector Spaces
3.1 Definitions and Examples

Modules are a generalization of the vector spaces of linear algebra in which
the "scalars" are allowed to be from an arbitrary ring, rather than a field.
This rather modest weakening of the axioms is quite far reaching, including,
for example, the theory of rings and ideals and the theory of abelian groups
as special cases.
(1.1) Definition. Let R be an arbitrary ring with identity (not necessarily

commutative).
(1) A left R-module (or left module over R) is an abelian group M together
with a scalar multiplication map
:RxM-M
that satisfy the following axioms (as is customary we will write am in
place of (a, m) for the scalar multiplication of m E 1LI by a E R). In
these axioms, a, b are arbitrary elements of R and m, n are arbitrary
elements of M.
(ai)a(m + n) = am + an.
(bi)(a + b)m = am + bm.
(ci) (ab)m = a(bm).
(di)lm = m.
(2) A right R-module (or right module over R) is an abelian group M
together with a scalar multiplication map
:MxR-+M
that satisfy the following axioms (again a, b are arbitrary elements of
R and m, n are arbitrary elements of M).
(&)(m + n)a = ma + na.
(b,)m(a + b) = ma + mb.
108 Chapter3. Modules and Vector Spaces
(cr)m(ab) = (ma)b.
(dr)m1 = m.
(1.2) Remarks.
(1) If R is a commutative ring then any left R-module also has the struc-
ture of a right R-module by defining mr = rm. The only axiom that
requires a check is axiom (cr). But
m(ab) = (ab)m = (ba)m = b(am) = b(ma) = (ma)b.
(2) More generally, if the ring R has an antiautomorphism (that is, an
additive homomorphism 0: R -' R such that 0(ab) = m(b)4(a)) then
any left R-module has the structure of a right R-module by defining
ma = ¢(a)m. Again, the only axiom that needs checking is axiom (cr):
(ma)b = .(b)(ma)
= O(b)(O(a)m)
= (m(b)O(a))m
= 0(ab)m
= m(ab).
An example of this situation occurs for the group ring R(G) where R
is a ring with identity and G is a group (see Example 2.1.10 (15)). In
this case the antiautomorphism is given by
-O(Ea99) _ a99-'
9EG 9EG
We leave it as an exercise to check that : R(G) - R(G) is an

antiautomorphism. Thus any left R(G)-module M is automatically a
right R(G)-module.
(3) Let R be an arbitrary ring and let R°P ("op" for opposite) be the
ring whose elements are the elements of R, whose addition agrees with
that of R, but whose multiplication - is given by a b = ba (where
the multiplication on the right-hand side of this equation is that of
R). Then any left R-module is naturally a right R°P-module (and vice-
versa). In fact, if M is a left R-module, define a right multiplication
of elements of R°P (which are the same as elements of R) on M by
am. As in Remark 1.2 (1), the only axiom that requires checking
is axiom (Cr). But
The theories of left R-modules and right R-modules are entirely par-
allel, and so, to avoid doing everything twice, we must choose to work on
3.1 Definitions and Examples 109
one side or the other. Thus, we shall work primarily with left R-modules
unless explicitly indicated otherwise and we will define an R-module (or
module over R) to be a left R -module. (Of course, if R is commutative, Re-
mark 1.2 (1) shows there is no difference between left and right R-modules.)
Applications of module theory to the theory of group representations will,
however, necessitate the use of both left and right modules over noncommu-
tative rings. Before presenting a collection of examples some more notation
will be introduced.
(1.3) Definition. Let R be a ring and let M, N be R-module8. A function

f : M -+ N is an R-module homomorpbism if
(1) A MI + m2) = A MO + P M2) for all ml, m2 E M, and
(2) f (am) = a f (m) for all a E R and m E M.
The set of all R-module homomorphisms from M to N will be de-

noted HomR(M, N). In case M = N we will usually write EndR(M) rather
than HomR(M, M); elements of EndR(M) are called endomorphisms. If
f E EndR(M) is invertible, then it is called an automorphism of M. The
group of all R-module automorphisms of M is denoted AutR(M) (Aut(M)
if R is implicit). If f E HomR(M, N) then we define Ker(f) C M and
Im(f) C N to be the kernel and image of f considered as an abelian group
homomorphism.
(1.4) Definition.
(1) Let F be a field. Then an F-module V is called a vector space over F.
(2) If V and W are vector spaces over the field F then a linear transfor-
mation from V to W is an F-module homomorphism from V to W.
(1.5) Examples.
(1) Let G be any abelian group and let g E G. If n E Z then define the
scalar multiplication ng by
(n terms) if n > 0,
ng= 0 ifn=0,
(-g) + + (-g) (-n terms) ifn < 0.
Using this scalar multiplication G is a Z-module. Furthermore, if G
and H are abelian groups and f : G -a H is a group homomorphism,
then f is also a Z-module homomorphism since (if n > 0)
f (ng) = f (9 + ... + 9) = f (9) + ... + f(9) = nf(9)

and f(-g) = -f(g).
110 Chapter 3. Modules and Vector Spaces
(2) Let R be an arbitrary ring. Then R" is both a left and a right R-module
via the scalar multiplications
a(bl, ... b") = (abi, ... ,ab")
and
(bi, ... , b" )a = (bia, ... , b"a).
(3) Let R be an arbitrary ring. Then the set of matrices M,,,,"(R) is both
a left and a right R-module via left and right scalar multiplication of
matrices, i.e.,
ent;j (aA) = a ent,J(A)
and
ent,3(Aa) = (ent,j(A))a.
(4) As a generalization of the above example, the matrix multiplication
maps
Mm(R) x M,,,,"(R) Mm."(R)

(A, B) .-- AB
and
Mm,"(R) x M"(R) Mm,"(R)
(A, B) - AB
make M,,,,"(R) into a left Mm(R)-module and a right M"(R)-module.
(5) If R is a ring then a left ideal I C R is a left R-module, while a right
ideal J C R is a right R-module. In both cases the scalar multiplication
is just the multiplication of the ring R.
(6) If R is a ring and I C R is an ideal then the quotient ring R/I is both
a left R-module and a right R-module via the multiplication maps
(a, b + I) ab + I
and
R/IxR-+R/I
(a+I,b).--sab+l.
(7) M is defined to be an R-algebra if M is both an R-module and a ring,
with the ring addition being the same as the module addition, and the
multiplication on M and the scalar multiplication by R satisfying the
following identity: For every r E R, m1, m2 E M,
(1.1) r(mlm2) = (rmi)m2 = ml(rm2).

3.1 Definitions and Examples 111
For example, every ring is a Z-algebra, and if R is a commutative ring,

then R is an R-algebra. Let R and S be rings and let 4 : R -+ S
be a ring homomorphism with Im(m) C C(S) = (a E S : ab = ba
for all b E S), the center of S. If M is an S-module, then M is also
an R-module using the scalar multiplication am = (4(a))m for all
a E R and m E M. Since S itself is an S-module, it follows that S
is an R-module, and moreover, since Im(¢) C C(S), we conclude that
S is an R-algebra. As particular cases of this construction, if R is a
commutative ring, then the polynomial ring R[X] and the matrix ring
M are R-modules then HomR(M, N) is an abelian group via

the operation (f + g)(m) = f (m) + g(m). However, if we try to make
HomR(M, N) into an R-module in the natural way by defining of by
the formula (a f)(m) = a(f (m)) we find that the function a f need not
be an R-module homomorphism unless R is a commutative ring. To
see this, note that
(af)(rm) = a(f (rm)) = a(r(f (m))) = ar f (m).
This last expression is equal to r(a f) (m) = ra f (m) if R is a commu-

tative ring, but not necessarily otherwise. Thus, if R is a commutative
ring, then we may consider HomR(M, N) as an R-module for all M,
N, while if R is not commutative then HomR(M, N) is only an abelian
group. Since EndR(M) is also a ring using composition of R-module
homomorphisms as the multiplication, and since there is a ring ho-
momorphism 0 : R - EndR(M) defined by O(a) = a 1M where IM
denotes the identity homomorphism of M, it follows from Example 1.5
(7) that EndR(M) is an R-algebra if R is a commutative ring.
(9) If G is an abelian group, then Homz(Z, G) G. To see this, define
$ : Homz(Z, G) - G by 4i(f) = f (1). We leave it as an exercise to
check that 4) is an isomorphism of Z-modules.
(10) Generalizing Example 1.5 (9), if M is an R-module then
HomR(R,M)^='M
as Z-modules via the map 4) : HomR(R, M) -+ M where 0(f) = f (1).
(11) Let R be a commutative ring, let M be an R-module, and let S C
EndR(M) be a subring. (Recall from Example 1.5 (8) that EndR(M)
is a ring, in fact, an R algebra.) Then M is an S-module by means of
the scalar multiplication map S x M M defined by (f, m) '--' f (m).
(12) As an important special case of Example 1.5 (11), let T E EndR(M)
and define a ring homomorphism 0 : R[X] EndR(M) by sending
X to T and a E R to alM. (See the polynomial substitution theorem
(Theorem 2.4.1).) Thus, if
f (X) =ao+a1X +...+aX"

then
O(f(X))
We will denote O(f (X)) by the symbol f (T) and we let Im(O) = R[T].
That is, R[T] is the subring of EndR(M) consisting of "polynomials"
in T. Then M is an R[TJ module by means of the multiplication
f (T)m = f (T)(m).
Using the homomorphism 0: R(X) -, R[T] we see that M is an R[X]-
module using the scalar multiplication
f (X)m = f (T)(m).
This example is an extremely important one. It provides the basis for
applying the theory of modules over principal ideal domains to the
study of linear transformations; it will be developed fully in Section
4.4.
(13) We will present a concrete example of the situation presented in Ex-
ample 1.5 (12). Let F be a field and define a linear transformation
T : F2 -+ F2 by T(ul, u2) = (u2, 0). Then T2 = 0, so if f (X) =
ao + a1X + + amXm E F[XJ, it follows that f (T) = ao1Fa + a1T.
Therefore the scalar multiplication f (X )u for u E F2 is given by
f(X) . (U1, U2) = f(T)(u1,u2)

= (ao1F2 +a,T)(u1,u2)
= (aoui +a1u2,aou2)
3.2 Submodules and Quotient Modules

Let R be a ring and M an R-module. A subset N C M is said to be
a submodule (or R-submodule) of M if N is a subgroup of the additive
group of M that is also an R-module using the scalar multiplication on
M. What this means, of course, is that N is a submodule of M if it is a
subgroup of M that is closed under scalar multiplication. These conditions
can be expressed as follows.
(2.1) Lemma. If M is an R-module and N is a nonempty subset of M,

then N is an R-submodule of M if and only if amt + bm2 E N for all
m1, m2 N and a, bE R.
Proof. Exercise.
3.2 Submodules and Quotient Modules 113
If F is a field and V is a vector space over F, then an F-submodule of

V is called a linear subspace of V.
(2.2) Examples.
If R is any ring then the R-submodules of the R-module R are precisely
the left ideals of the ring R.
If G is any abelian group then G is a Z-module and the Z-submodules
of G are just the subgroups of G.
Let f : M - N be an R-module homomorphism. Then Ker(f) C M
and Im(f) C N are R-submodules (exercise).
Continuing with Example 1.5 (12), let V be a vector space over a
field F and let T E EndF(V) be a fixed linear transformation. Let VT
denote V with the FIX)-module structure determined by the linear
transformation T. Then a subset W C V is an F[X]-submodule of the
module VT if and only if W is a linear subspace of V and T (W) C W,
i.e., W must be a T-invariant subspace of V. To see this, note that
X w = T(w), and if a E F, then a w = aw-that is to say, the
action of the constant polynomial a E F[X] on V is just ordinary
scalar multiplication, while the action of the polynomial X on V is
the action of T on V. Thus, an F[X]-submodule of VT must be a T-
invariant subspace of V. Conversely, if W is a linear subspace of V
such that T(W) C W then Tm(W) C W for all m > 1. Hence, if
f (X) E F[X] and W E W then f (X) w = f (T)(w) E W so that W is
closed under scalar multiplication and thus W is an F[X]-submodule
of V.
(2.3) Lemma. Let M be an R-module and let (NQ}QEA be a family of sub-

modules of M. Then N Z n0EA N. is a submodule of M.
Proof. Exercise. 0
We now consider quotient modules and the noether isomorphism the-
orems. Let M be an R-module and let N C_ M be a submodule. Then N
is a subgroup of the abelian group M, so we can form the quotient group
M/N. Define a scalar multiplication map on the abelian group M/N by
a(m + N) = am + N for all a E R, m + N E M/N. Since N is an R-
submodule of M, this map is well defined. Indeed, if m + N = m' + N then
m-m' E N so that am-am'= a(m-m') E N so that am+N = am'+N.
The resulting R-module M/N is called the quotient module of M with re-
spect to the submodule N. The noether isomorphism theorems, which we
have seen previously for groups and rings, then have direct analogues for
R-modules.
(2.4) Theorem. (First isomorphism theorem) Let M and N be modules

over the ring R and let f : M N be an R-module homomorphism. Then
Im(f).
Proof. Let K = Ker(f). From Theorem 1.3.10 we know that f : M/K -

Im(f) defined by f (m+K) = f (m) is a well-defined isomorphism of abelian
groups. It only remains to check that f is an R-module homomorphism. But
f(a(m+K)) = f(am+K) = f(am) = af(m) = af(m+K) for all m E M
and a E R, so we are done. 0
(2.5) Theorem. (Second isomorphism theorem) Let M be an R-module and

let N and P be submodules. Then there is an isomorphism of R-modules
(N+P)/PAN/(NnP).
Proof. Let rr : M , M/P be the natural projection map and let rro be
the restriction of rr to N. Then ao is an R-module homomorphism with
Ker(rro) = N n P and Im(iro) = (N + P)/P. The result then follows from
the first isomorphism theorem. 0
(2.6) Theorem. (Third isomorphism theorem) Let M be an R-module and

let N and P be submodules of M with P C N. Then
111/N = (M/P)/(N/P).
Proof. Define f : Al/P -. MIN by f (m+ P) = m+N. This is a well-defined

R-module homomorphism and
Ker(f)={m +P:m+N=N}={m +P:mEN}=NIP.

The result then follows from the first isomorphism theorem (Theorem 2.4).
0
(2.7) Theorem. (Correspondence theorem) Let Al be an R-module, N a

submodule, and A : M - MIN the natural projection map. Then the func-
tion P -- P/N defines a one-to-one correspondence between the set of all
submodules of M that contain N and the set of all submodules of M/N.
Proof. Exercise. 0
(2.8) Definition. If S is a subset of an R-module M then (S) will denote
the intersection of all the submodules of M that contain S. This is called
the submodule of M generated by S, while the elements of S are called
generators of (S).
Thus, (S) is a submodule of M that contains S and it is contained in

every submodule of M that contains S, i.e., (S) is the smallest submodule
of M containing S. If S = {x1, ... ,xn} we will usually write (xi, ... ,xn)
rather than ({x1, ... , for the submodule generated by S. There is the
following simple description of (S).
(2.9) Lemma. Let M be an R-module and let S C M. If S = 0 then (S) =

{0},while (S)={E 1a;si:nEN,a;ER, s,ES,1<i<n}ifS#0.
Proof. Exercise.
(2.10) Definition. We say that the R-module M is finitely generated if

M = (S) for some finite subset S of M. M is cyclic if M = (m) for
some element m E M. If M is finitely generated then let µ(M) denote the
minimal number of generators of M. If M is not finitely generated, then let
µ(M) = oo. We will call u(M) the rank of M.
(2.11) Remarks.
(1) We have uc({O}) = 0 by Lemma 2.9 (1), and M # {0} is cyclic if and
only if µ(M) = 1.
(2) The concept of cyclic R-module generalizes the concept of cyclic group.
Thus an abelian group G is cyclic (as an abelian group) if and only if
it is a cyclic Z-module.
(3) If R is a PID, then any R-submodule M of R is an ideal, so µ(M) = 1.
(4) For a general ring R, it is not necessarily the case that if N is a sub-
module of the R-module M, then µ(N) < µ(M). For example, if R is
a polynomial ring over a field F in k variables, M= R, and N C M
is the submodule consisting of polynomials whose constant term is 0,
then u(M) = 1 but µ(N) = k. Note that this holds even if k = oo. We
shall prove in Corollary 6.4 that this phenomenon cannot occur if R is
a PID. Also see Remark 6.5.
If M is a finitely generated R-module and N is any submodule, then

M/N is clearly finitely generated, and in fact, µ(M/N) < µ(M) since
the image in M/N of any generating set of M is a generating set of M/N.
There is also the following result, which is frequently useful for constructing
arguments using induction on µ(M).
(2.12) Proposition. Suppose M is an R-module and N is a submodule. If

N and M/N are finitely generated, then so is M and
µ(M) < µ(N) + µ(M/N).
Proof. Let S = {x1, ... , xk} C N be a minimal generating set for N and if
rr : M -+ M/N is the natural projection map, choose T = {yl, ... , yt} C M
so that {rr(yl), ... ,7r(ye)} is a minimal generating set for M/N. We claim
that S U T generates M so that u(M) < k + I = µ(N) + µ(M/N). To see
this suppose that x E M. Then rr(x) = al7r(y1) + + atrr(yt). Let y =
alyl+ +atyt E (T). Then ir(x-y) =O so that x-y E Ker(ir) = N = (S).

It follows that x = (x - y) + y E (S U T), and the proof is complete. 0
(2.13) Definition. If {NQ}QEA is a family of R-submodules of M, then the
submodule generated by {NQ}QEA is (UQEA NQ). This is just the set of all
sums nQ, + + n(kk where nQ, E NQ, . Instead of (UQEA N") , we will use
the notation EQEA NQ; if the index set A is finite, e.g., A = {1, ... , m},
we will write N1 + + Nn for the submodule generated by N1, ... , Nn.
(2.14) Definition. If R is a ring, M is an R-module, and X is a subset of

M, then the annihilator of X, denoted Ann(X), is defined by
Ann(X)={aE R:ax=O forallxEX}.
It is easy to check that Ann(X) is a left ideal of R, and furthermore,

if X = N is a submodule of M, then Ann(N) is an ideal of R. If R is
commutative and N = (x) is a cyclic submodule of M with generator x,
then
Ann(N)={aER:ax=O}.
This fact is not true if the ring R is not commutative. As an example, let
R = M,, (R) = M and let x = Ell be the matrix with a 1 in the 11 position
and 0 elsewhere. It is a simple exercise to check that Ann(Ell) consists of
all matrices with first column 0, while Ann((E11)) = (0).
If R is commutative and N is cyclic with generator x then we will
usually write Ann(x) rather than Ann((x)). In this situation, the ideal
Ann(x) is frequently called the order ideal of x. To see why, consider the
example of an abelian group G and an element g E G. Then G is a Z-module
and
Ann(g)={nEZ:ng=0}
_ (p)
where p = o(g) if o(g) < oo and p = 0 if (g) is infinite cyclic.
Example. Let F be a field, V a vector space, T E EndF(V) a linear trans-

formation, and let VT be the F[X] module determined by T (Example 1.5
(12)). If v E V then
Ann(v) = if (X) E F[X] : f (T)(v) = 0}.
Note that this is a principal ideal (g(X)) since F[X] is a PID.
(2.15) Proposition. Let R be a ring and let M = (m) be a cyclic R-module.

Then M R/ Ann(m).
Proof. The function f : R -+ M defined by f (a) = am is a surjective R-
module homomorphism with Ker(f) = Ann(m). The result follows by the
first isomorphism theorem. 0
(2.16) Corollary. If F is a field and M is a nonzero cyclic F-module then

Mn-iF.
Proof. A field has only the ideals {0} and F, and 1 m = m t 0 if m 36 0
is a generator for M. Thus, Ann(m) 36 F, so it must be {0}.
If M is an R-module and I C R is an ideal then we can define the

product of I and M by
IM={>aimi:nEN,aiEI,miEM}.
,_ 1 JJJ
The set IM is easily checked to be a submodule of M. The product IM

is a generalization of the concept of product of ideals. If R is commutative
and I C Ann(M) then there is a map
R/IxM -+M
defined by (a + I )m = am. To see that this map is well defined, suppose
that a + I = b + I. Then a - b E I C Ann(M) so that (a - b)m = 0, i.e.,
am = bm. Therefore, whenever an ideal I C Ann(M), M is also an R/I
module. A particular case where this occurs is if N = M/IM where I is any
ideal of R. Then certainly I C Ann(N) so that M/IM is an R/I-module.
(2.17) Definition. Let R be an integral domain and let M be an R-module.

We say that an element x E M is a torsion element if Ann(x) ,-f {0}. Thus
an element x E M is torsion if and only if there is an a # 0 E R such
that ax = 0. Let M, be the set of torsion elements of M. M is said to be
torsion-free if M, = {0}, and M is a torsion module if M = M,.
(2.18) Proposition. Let R be an integral domain and let M be an R-module.

(1) M, is a submodule of M, called the torsion submodule.
(2) M/M, is torsion-free.
Proof. (1) Let x, y E M, and let c, d E R. There are a 34 0, b 36 0 E R such

that ax = 0 and by = 0. Since R is an integral domain, ab # 0. Therefore,
ab(cx + dy) = bc(ax) + ad(by) = 0 so that cx + dy E M.
(2) Suppose that a 0 0 E R and a(x + M,) = 0 E (M/M,),. Then
ax there is ab00ERwith (ba)x=b(ax)=0.Since ba54 0, it
follows that x E M i.e., x + M, = 0 E M/M,.
(2.19) Examples.
(1) If G is an abelian group then the torsion Z-submodule of G is the
set of all elements of G of finite order. Thus, G = G, means that
every element of 0 is of finite order. In particular, any finite abelian
group is torsion. The converse is not true. For a concrete example, take
G = Q/Z. Then IGI = oo, but every element of Q/Z has finite order
since q(p/q + Z) = p + Z = 0 E Q/Z. Thus (Q/Z),. = Q/Z.
(2) An abelian group is torsion-free if it has no elements of finite order
other than 0. As an example, take G = Z" for any natural number n.
Another useful example to keep in mind is the additive group Q.
(3) Let V = F2 and consider the linear transformation T : F2 -+ F2
defined by T(ul, u2) = (u2i 0). See Example 1.5 (13). Then the F[X)
module VT determined by T is a torsion module. In fact Ann(VT) =
(X2). To see this, note that T2 = 0, so X2 u = 0 for all u E V. Thus,
(X2) C Ann(VT). The only ideals of F[XI properly containing (X2)
are (X) and the whole ring F[X], but X 0 Ann(VT) since X (0, 1) _
(1, 0) 54 (0, 0). Therefore, Ann(VT) _ (X2).
The following two observations are frequently useful; the proofs are left
as exercises:
(2.20) Proposition. Let R be an integral domain and let M be a finitely gen-

erated torsion R-module. Then Ann(M) 34 (0). In fact, if M = (x1i ... , xn)
then
Ann(M) = Ann(x1) n . . . n Ann(xn) 34 (0).
Proof. Exercise.
(2.21) Proposition. Let F be afield and let V be a vector space over F, i.e.,
an F-module. Then V is torsion-free.
Proof. Exercise.
3.3 Direct Sums, Exact Sequences, and Hom

Let M1, ... , AIn be a finite collection of R-modules. Then the cartesian
product set M1 x . x Mn can be made into an R-module by the operations
(x1, ... , xn) + (y1, ... , yn) _ (xl +Y1, ... , xn +Y.)
a(x1i ... , xn) = (axi, ... , axn)
where the 0 element is, of course, (0, ... , 0). The R-module thus con-
structed is called the direct sum of M1, ... , Mn and is denoted
n
M1 ®...®Mn (or ®M;).
=1
3.3 Direct Sums, Exact Sequences, and Horn 119
The direct sum has an important homomorphism property, which, in

fact, can be used to characterize direct sums. To describe this, suppose that
fi : Mi -+ N are R-module homomorphisms. Then there is a map
f:Ml®...ED Mn-+N
defined by
n
f(xi, ... ,x,) _ fi(xi)-
i=1
We leave it as an exercise to check that f is an R-module homomorphism.
Now consider the question of when a module M is isomorphic to the
direct sum of finitely many submodules. This result should be compared
with Proposition 1.6.3 concerning internal direct products of groups.
(3.1) Theorem. Let M be an R-module and let Ml, ... , M,, be submodules
of M such that
(1) M=M1+ +Mn, and
(2) for 1 < i < n,
Then
M1®... ®Mn.
Proof. Let fi : Mi -+ M be the inclusion map, that is, fi(x) = x for all
x E Mi and define
f:Ml®...(D MMn-+M
by
f(xl,...,xn) = xl + ... + xn.
f is an R-module homomorphism and it follows from condition (1) that f is

surjective. Now suppose that (x1, ... , xn) E Ker(f ). Then x1 +- +xn = 0
so that for 1 < i < n we have
xi = -(XI + ... + xi_1 + xi+l + ... + xn)-
Therefore,
xi E =0
so that (x1, ... ,xn) = 0 and f is an isomorphism. O
Our primary emphasis will be on the finite direct sums of modules just
constructed, but for the purpose of allowing for potentially infinite rank
free modules, it is convenient to have available the concept of an arbitrary
direct sum of R-modules. This is described as follows. Let {Mj)jEj be
a family of R-modules indexed by the (possibly infinite) set J. Then the

cartesian product set fl J Mj is the set of all the indexed sets of elements
(xj)jEJ where xj is chosen from Mi. This set is made into an R-module by
the coordinate-wise addition and scalar multiplication of elements. More
precisely, we define
(xj)jEJ'+' (yj)JEJ = (x, +Yj)jE

a(xj)jEJ = (axj)jEJ
For each k E J there is an R-rnodule homomorphism ak : IIjEJ 'V'i - Mk
defined by 7rk((xj)jEJ) = xj, that is, irk picks out the element of the indexed
set (xj)JEJ that is indexed by k. We define the direct sum of the indexed
family { Mj }jEJ of R-modules to be the following submodule ®jEJ Mj of
jEJ Mj:
®Mj = {(xj)jEJ : xj = 0 except for finitely many indices j E J}.

jEJ
It is easy to check that ®jEJMj is a submodule of IIjEJ Mj.

To get a feeling for the difference between direct sums and direct prod-
ucts when the index set is infinite, note that the polynomial ring R[XJ, as an
R-module (ignoring the multiplicative structure), is just a countable direct
sum of copies of R, in fact, the nth copy of R is indexed by the monomial
X". However, the formal power series ring R[[X]], as an R-module, is just a
countable direct product of copies of R. Again, the nth copy of R is indexed
by the monomial X". Each element of the polynomial ring has only finitely
many monomials with nonzero coefficients, while an element of the formal
power series ring may have all coefficients nonzero.
The homomorphism property of the finite direct sum of R-modules
extends in a natural way to arbitrary direct sums. That is, suppose that
N is an arbitrary R-module and that for each j E J there is an R-module
homomorphism fj : Mj N. Then there is a map f : ®jEJMj - N
defined by f ((xj) jE J) = >jE J f j (x j ). Note that this sum can be considered
as a well-defined finite sum since xj = 0 except for finitely many indices j E
J. (Note that this construction does not work for infinite direct products.)
We leave it as an exercise to check that f is an R-module homomorphism.
The characterization of when an R-module M is isomorphic to the
direct sum of submodules is essentially the same as the characterization
provided in Theorem 3.1. We state the result, but the verification is left as
an exercise.
(3.2) Theorem. Let M be an R-module and let {Mj}jEJ be a family of

submodules such that
(1) M = >F,EJ
-,E.,Mj = (UjEJ Mj), and
(2) Mk n Mj = {0} for every k E J.
3.3 Direct Sums, Exact Sequences, and Hom 121
Then
MMi.
jEJ
Proof. Exercise. 0
(3.3) Definition. If M is an R-module and MI C M is a submodule, we say

that MI is a direct summand of M, or is complemented in M, if there is
a submodule M 2 C M such that M ? 5 MI ®M2.
(3.4) Example. Let R = Z and M = Z e. If MI = (p) then MI is not

complemented since MI is the only subgroup of M of order p, so condition
(2) of Theorem 3.1 is impossible to satisfy.
The concept of exact sequences of R-modules and R-module homo-

morphisms and their relation to direct summands is a useful tool to have
available in the study of modules. We start by defining exact sequences of
R-modules.
(3.5) Definition. Let R be a ring. A sequence of R-modules and R-module

homomorphisms
is said to be exact at Mi if Im(fi) = Ker(fi+I ). The sequence is said to be

exact if it is exact at each Mi.
As particular cases of this definition note that

(1) 0 -+ MI --f-+ M is exact if and only if f is injective,
(2) M -+ M2 - 0 is exact if and only if g is surjective, and
(3) the sequence
(3.1) 0 - . MI _ M 9-+ M2'0

is exact if and only if f is injective, g is surjective, and Im(f) = Ker(g).
Note that the first isomorphism theorem (Theorem 2.4) then shows
that M2 ?t M/ Im(f). M/ Im(f) is called the cokernel of f and it is
denoted Coker(f).
(3.6) Definition.
(1) The sequence (3.1), if exact, is said to be a short exact sequence.
(2) The sequence (3.1) is said to be a split exact sequence (or just split)
if it is exact and if Im(f) = Ker(g) is a direct summand of M.
In the language of exact sequences, Proposition 2.12 can be stated as

follows:
(3.7) Proposition. Let 0 -. M1 -+ M -i M2 - 0 be a short exact

sequence of R-modules. If Ml and M2 are finitely generated, then so is M,
and moreover,
µ(M) <_ µ(M1) + µ(M2)
Proof 0
(3.8) Example. Let p and q be distinct primes. Then we have short exact
sequences
(3.2) 0 - ZP Zpq Zq 0
and
(3.3) 0,Z,- Zp2-ZP- 0

where 0(m) = qm E Zpg, f (m) = pm E Zp2, and 0 and g are the canonical
projection maps. Exact sequence (3.2) is split exact while exact sequence
(3.3) is not split exact. Both of these observations are easy consequences of
the material on cyclic groups from Chapter 1; details are left as an exercise.
There is the following useful criterion for a short exact sequence to be

split exact.
(3.9) Theorem. If
(3.4) O- M1 -I.M 9M2--_0

is a short exact sequence of R-modules, then the following are equivalent:
(1) Them exists a homomorphism a : M M, such that a o f = 1 s,, .
(2) There exists a homomorphism 3: M2 -. M such that g o R = 1nt,
(3) The sequence (3.4) is split exact.
If these equivalent conditions hold then
M = Im(f) Eli Ker(a)
Ker(g) EH Im(p)
M1 e M2.
The homomorphisms a and ,3 are said to split the exact sequence (3.4)
or be a splitting.
Proof. Suppose that (1) is satisfied and let x E M. Then
a(x - f(a(x))) = a(x) - (a o f)((k(x)) = 0
since a o f Therefore, x - f(a(x)) E Ker(a) so that
Al - Ker(a) + Im(f ).
Now suppose that f (y) = x E Ker(a) fl Im(f). Then

0 = a(x) = a(f(y)) = y,
and therefore, x = f (y) = 0. Theorem 3.1 then shows that
M = Im(f) ® Ker(a).
Define,Q : M2 -- M by
(3.5) /3(u) = v - f(a(v))
where g(v) = u. Since g is surjective, there is such a v E M, but it may

be possible to write u = g(v) for more than one choice of v. Therefore, we
must verify that /3 is well defined. Suppose that g(v) = u = g(v'). Then
V - v' E Ker(g) = Im(f) so that
(v - f(a(v))) - (v' - f(a(v'))) = (v - v') + (f(a(v') - f(a(v)))
E lm(f) n Ker(a)
= {0}.
We conclude that /3 is well defined. Since it is clear from the construction
of /3 that g o /3 = ln! we have verified that (1) implies (2) and that
M Im(f) ® Ker(a), i.e., that (3) holds.
The proof that (2) implies (1) and (3) is similar and is left as an
exercise.
Suppose that (3) holds, that is, M =' M' (D M" where M' = Ker(g) _
Im(f). Let 7r1 : M - M' and 7r2 : M M" be the projections, and
c : M" M be the inclusion. Note that 7r1 o f : M1 M' and g o t :
M" -y M2 are isomorphisms. Define a : M All by a = (7r1 o f)-1 o 7r1
and/3:M2-+Mby/3 to(got)-l. Then aof=1M, and go/=1412,
so (1) and (2) hold. 0
If M and N are R-modules, then the set HomR(M, N) of all R-module
homomorphisms f : M -' N is an abelian group under function addition.
According to Example 1.5 (8), HomR(M, N) is also an R-module provided
that R is a commutative ring. Recall that EndR(M) = HomR(M) denotes
the endomorphism ring of the R-module M, and the ring multiplication is
composition of homomorphisms. Example 1.5 (8) shows that EndR(M) is
an R-algebra if the ring R is commutative. Example 1.5 (10) shows that
HomR(R, M) ^_' M for any R-module M.
Now consider R-modules M, M1, N, and N1, and let : N , N1,
i(i : M -. M1 be R-module homomorphisms. Then there are functions
0.: HomR(M, N) HomR(M, N1)

and
0*: HomR(Ml, N) -' HomR(M, N)
defined by
0.(f) = 0 of for all f E HomR(M, N)
and
tP* (g) = go 10 for all g E HomR(Ml, N).
It is straightforward to check that 0.(f +g) = 0. (f) +0. (g) and ?P* (f +g) _
Ii' (f) + t/5' (g) for appropriate f and g. That is, 0. and +f are homomor-
phisms of abelian groups, and if R is commutative, then they are also
R-module homomorphisms.
Given a sequence of R-modules and R-module homomorphisms
(3.6) lVij-1 .m`.

A-+: M,+i ...
and an R-module N, then HomR( , N) and HomR(N, ) produce two
sequences of abelian groups (R-modules if R is commutative):
(m.).
(3.7) HomR(N, Mi-1) HomR(N, M:)
HomR(N, M1+1) - .. .
and
(00,
(3.8) Hom M HomR(Mi, N)
i01+i.
HomR(Ms+1, N)
A natural question is to what extent does exactness of sequence (3.6)
imply exactness of sequences (3.7) and (3.8). One result along these lines
is the following.
(3.10) Theorem. Let
(3.9)
be a sequence of R-modules and R-module homomorphisms. Then the se-

quence (3.9) is exact if and only if the sequence
(3.10) 0 -. HomR(N, M1) m'+ HomR(N, M) 1P' HomR(N, M2)

is an exact sequence of Z-modules for all R-modules N.
if
(3.11) M1 0
is a sequence of R-modules and R-module homomorphisms, then the se-

quence (3.11) is exact if and only if the sequence
(3.12) 0 HomR(M2, N) --+ HomR(M, N) m . HomR(Ml, N)

is an exact sequence of Z-modules for all R-modules N.
Proof. Assume that sequence (3.9) is exact and let N be an arbitrary R-

module. Suppose that f E HomR(N, M) and -0.(f) = 0. Then
0 = 0 o f (x) = g5(f (x))
for all x E N. But 0 is injective, so f (x) = 0 for all x E N. That is, f = 0,

and hence, 0. is injective.
Since o 0 = 0 (because sequence (3.9) is exact at M), it follows that
IMO-W) =+G-0.(f) =0000f =0

for all f E HomR(N, M). Thus Im(q.) C Ker(t/'.). It remains to check
the other inclusion. Suppose that g E HomR(N, M) with t1.(g) = 0, i.e.,
0(g(x)) = 0 for all x E N. Since Ker(tP) = Im(¢), for each x E N, we
may write g(x) = 4(y) with y E Ml. Since 0 is injective, y is uniquely
determined by the equation g(x) = 0(y). Thus it is possible to define a
function f : N - Ml by f (x) = y whenever g(x) = ¢(y). We leave it as an
exercise to check that f is an R-module homomorphism. Since 0.(f) = g,
we conclude that Ker(t/'.) = Im(O.) so that sequence (3.10) is exact.
Exactness of sequence (3.12) is a similar argument, which is left as an
exercise.
Conversely, assume that sequence (3.10) is exact for all R-modules
N. Then 0. is injective for all R-modules N. Then letting N = Ker(O)
and t : N - Ml be the inclusion, we see that 0.(t) = 0 o t = 0. Since
0.: HomR(N, Ml) --- HomR(N, M) is injective, we see that t = 0, i.e.,
N = (0). Thus, 0 is injective.
Now letting N = MI we see that
0= (0.00.)(1m.) =V,00.
Thus Im(O) C Ker(O). Now let N = Ker(t(i) and let t : N -+ M be the
inclusion. Since tp. (t) = 7P o t = 0, exactness of Equation (3.10) implies that
t = ¢.(a) for some a E HomR(N, MI). Thus,
Im(O) Im(t) = N = Ker(tP),
and we conclude that sequence (3.9) is exact.
Again, exactness of sequence (3.11) is left as an exercise.
Note that, even if
0--.MI-.M 1P+M2-i0
is a short exact sequence, the sequences (3.10) and (3.12) need not be short
exact, i.e., neither tG. or 0' need be surjective. Following are some examples
to illustrate this.
(3.11) Example. Consider the following short exact sequence of Z-modules:
(3.13) 0- Z--0+ Z_4Z,n--+ 0

where 4(i) = mi and tai is the canonical projection map. If N = Z,, then
sequence (3.12) becomes
0 ---+ Homz(Z,n, Z,,) ----+ Homz(Z, Zn) 0' Homz(Z, Zn),

which, by Example 1.5 (10), becomes
0 --+ Homz(Zm, Zn) ----* Z" ----+ Z"

so that
Homz(Z,n, Zn) = Ker(O*).
Let d = gcd(m, n), and write m = m'd, n = n'd. Let f E Homz(Z, Zn).
Then, clearly, 0* (f) = 0 if and only if 0* (f) (1) = 0. But
0*(f)(1) = f(m 1) = mf(1) = m'df(1).
Since m' is relatively prime to n, we have m'df (1) = 0 if and only if df(1) _
0, and this is true if and only if f (1) E n'Zn. Hence, Ker(O') = n'Zn Zd,
i.e.,
(3.14) Homz(Zm, Zn) = Zd.

This example also shows that even if
0,Af,-+M- M2- 0
is exact, the sequences (3.10) and (3.12) are not, in general, part of short
exact sequences. For simplicity, take m = n. Then sequence (3.12) becomes
(3.15) 0-4 Zn-+ Zn -4Zn

with 0' = 0 so that 0' is not surjective, while sequence (3.10) becomes
V-4
!3.16) 0 -' Homz(Z,,, Z) -+ Homz(Z,,, Z) Homz(Zn, Zn).
Since Homz(Zn, Z) = 0 and Homz(Zn, Zn) Zn, sequence (3.16) becomes
0-+0--a0 ±- +Zn
and ii. is certainly not surjective.
These examples show that Theorem 3.10 is the best statement that
can be made in complete generality concerning preservation of exactness
under application of HomR. There is, however, the following criterion for
the preservation of short exact sequences under Hom:
3.3 Direct Sums, Exact Sequences, and Horn 127
(3.12) Theorem. Let N be an arbitrary R-module. If
(3.17) 0-4 M1-0- M 0-+ M2

is a split short exact sequence of R-modules, then
(3.18) 0 -. HomR(N, M1) 0.i HomR(N, M) - HomR(N, M2) 0
and
(3.19) 0 HomR(M2, N) HomR(M, N) HorR(M1, N) 0
are split short exact sequences of abelian groups (R-modules if R is com-

mutative).
Proof. We will prove the split exactness of sequence (3.18); (3.19) is similar
and it is left as an exercise. Given Theorem 3.10, it is only necessary to
show that t&, is surjective and that there is a splitting for sequence (3.18).
Let /3 : M2 -+ M split the exact sequence (3.17) and let f E HomR(N, M2).
Then
00f)
= (WOQ)0f
0.aa.(f)=IM,,,
= (1M2) ° f
= (1HomR(N,Ma)) (f)
Thus, r/). o)3. = 1HomR(N,M2) so that r/i. is surjective and /i. is a splitting
of exact sequence (3.18).
(3.13) Corollary. Let M1, M2, and N be R-modules. Then

(3.20) HomR(N, M1 ® M2) 25 HomR(N, M1) ® HomR(N, M2)
and
(3.21) HomR(Ml (B M2, N) 25 HomR(M1, N) ED HorR(M2, N).
The isomorphisms are Z-module isomorphisms (R-module isomorphisms if
R is commutative).
Proof. Both isomorphisms follow by applying Theorems 3.12 and 3.9 to the
split exact sequence
0-+ M1 `+ 0
where c(m) = (m, 0) is the canonical injection and rr(ml, m2) = m2 is the
canonical projection.
(3.14) Remarks.
(1) Notice that isomorphism (3.20) is given explicitly by
't(f)=(r1of,7rzof)
where f E HomR(N, M1 ® M2) and Jri(ml, m2) = mi (for i = 1, 2);
while isomorphism (3.21) is given explicitly by
4'(f)=(fo4,foL2)
where f E HomR(Ml ® M2, N), Li : M1 - M1 ® M2 is given by
G1(m)=(m,0)andi2:M2- M1®M2is given by t2(m)=(0,m).
(2) Corollary 3.13 actually has a natural extension to arbitrary (not nec-
essarily finite) direct sums. We conclude this section by stating this
extension. The proof is left as an exercise for the reader.
(3.15) Proposition. Let {Mi},EI and {Nj}jEJ be indexed families (not

necessarily finite) of R-modules, and let M = ®,EIMi, N = ®jEJNj. Then
HomR(M, N) = [I(®HomR(Mi, Nj)).

iEI jEJ
Proof. Exercise. 0
3.4 Free Modules
(4.1) Definition. Let R be a ring and let M be an R-module. A subset S C M

is said to be R-linearly dependent if there exist distinct x1, ... , xn in S and
elements a1, ... , an of R, not all of which are 0, such that
A set that is not R-linearly dependent is said to be R-linearly independent.
When the ring R is implicit from the context, we will sometimes write
linearly dependent (or just dependent) and linearly independent (or just
independent) in place of the more cumbersome R-linearly dependent or
R-linearly independent. In case S contains only finitely many elements
x1, x2i ... , xn, we will sometimes say that x1, x2, ... , x, are R-linearly de-
pendent or R-linearly independent instead of saying that S = {x1, ... , xn}
is R-linearly dependent or R-linearly independent.
3.4 Flee Modules 129
(4.2) Remarks.
(1) To say that S C M is R-linearly independent means that whenever
there is an equation
=0
where x1 i ... , xn are distinct elements of S and a1, ... , an are in R,
then
a1 = =an =0.
(2) Any set S that contains a linearly dependent set is linearly dependent.
(3) Any subset of a linearly independent set S is linearly independent.
(4) Any set that contains 0 is linearly dependent since 1 -0 = 0.
(5) A set S C M is linearly independent if and only if every finite subset
of S is linearly independent.
(4.3) Definition. Let M be an R-module. A subset S of M is a basis of M

if S generates M as an R-module and if S is R-linearly independent. That
is, S C M is a basis if and only if M = {0}, in which case S = 0 is a basis,
or M 0 {0} and
(1) every x E M can be written as
x = a1x1 + ... + anxn

for some x1i ... , xn E S and a1, ... , an E R, and
(2) whenever there is an equation
a1x1 + ... + anxn = 0
where x1i ... , xn are distinct elements of S and a1, ... , an are in R,
then
a1 = ... = an = 0.
It is clear that conditions (1) and (2) in the definition of basis can be
replaced by the single condition:
(1') S C M is a basis of M 36 {0} if and only if every x E M can be written
uniquely as
x=alxl+...+anxn
foral,...,anERandx1,...,xnES.
(4.4) Definition. An R-module M is a free R-module if it has a basis.
(4.5) Remark. According to Theorem 3.2, to say that S = {xj }jE J is a

basis of M is equivalent to M being the direct sum of the family {Rx,)jEJ
of submodules of M, where Ann(xj) = {0} for all j E J. Moreover, if J is

any index set, then N = (DjE JRj, where Rj = R for all j E J, is a free R-
module with basis S = {ej}jEJ, where ej E N is defined by ej = (bjk)kEJ.
Here, bjk is the kronecker delta function, i.e., bjk = 1 E R whenever j = k
and bjk = 0 E R otherwise. N is said to be fine on the index set J.
(4.6) Examples.
(1) If R is a field then R-linear independence and R-linear dependence in
a vector space V over R are the same concepts used in linear algebra.
(2) R' is a free module with basis S = {el, ... , en} where
e; = (0,...,0,1,0,...,0)
with a 1 in the ith position.

(3) Mm,n(R) is a free R-module with basis
S={E,j:1<i<m,I<j<n}.
(4) The ring R[X] is a free R-module with basis {X" : n E Z+}. As in
Example 4.6 (2), R[XJ is also a free R[X]-module with basis {1}.
(5) If C is a finite abelian group then G is a Z-module, but no nonempty
subset of G is Z-linearly independent. Indeed, if g E G then IGI g = 0
but Cl I34 0. Therefore, finite abelian groups can never be free Z-
modules, except in the trivial case G = {0} when 0 is a basis.
(6) If R is a commutative ring and I C R is an ideal, then I is an R-
module. However, if I is not a principal ideal, then I is not free as an
R-module. Indeed, no generating set of I can be linearly independent
since the equation (-a2)al +ala2 = 0 is valid for any al, a2 E R.
(7) If M1 and M2 are free R-modules with bases Sl and S2 respectively,
then Ml ® M2 is a free R-module with basis S; U S2, where
Sj ={(x,0):iES1} and S2={(0,y):yeS2}.

(8) More generally, if {Mj}jEJ is a family of free R-modules and Sj C ltlj
is a basis of Mj for each j E J, then M = ®jEJMj is a free R-module
and S = UjEJS is a basis of M, where S' C M is defined by
Sj = {Sjo = (bjksjt)kEJ Sj, E Sj}.
Informally, Si consists of all elements of M that contain an element of

Sj in the jth component and 0 in all other components. This example
incorporates both Example 4.6 (7) and Example 4.6 (2).
Example 4.6 (5) can be generalized to the following fact.

3.4 Free Modules 131
(4.7) Lemma. Let M be an R-module where R is a commutative ring. Then

an element x E M is R-independent if and only if Ann(x) = {0}. In par-
ticular, an element a E R is an R-independent subset of the R-module R if
and only if a is not a zero divisor.
Proof. Exercise.
(4.8) Proposition. Let R be an integral domain and let M be a free R-

module. Then M is torsion-free.
Proof. Let M have a basis S = {x) }jEJ and let x E M,. Then ax = 0 for
some a 96 0 E R. Write x = /jEJ ajxj. Then
0 = ax = 1:(aaj)xj.
jEJ
Since S is a basis of M, it follows that aaj = 0 for all j E J, and since

a 34 0 and R is an integral domain, we conclude that aj = 0 for all j E J.
Therefore, x = 0, and hence, M, = (0) so that M is torsion-free.
The existence of a basis for an R-module M greatly facilitates the

construction of R-module homomorphisms from M to another R-module
N. In fact, there is the following important observation.
(4.9) Proposition. Let M be a free R-module with basis S, let N be any

R-module, and let h : S - N be any function. Then there is a unique
f E HomR(M, N) such that f Is = h.
Proof. Let S = {xj}jEJ. Then any x E M can be written uniquely as
x= jEJ ajxj where at most finitely many aj are not 0. Define f : M - N
by
f (x) = > ajh(xj)

jEJ
It is straightforward to check that f E HomR(M, N) and that f is = h.
Remark. The content of Proposition 4.9 is usually expressed as saying that

the value of a homomorphism can be arbitrarily assigned on a basis.
(4.10) Corollary. Suppose that M is a free R-module with basis S = {xj}jEJ.

Then
HomR(M, N) II Nj
jEJ
where N. = N for all j E J.

Proof. Define 4i : HomR(M, N) rjjEJ Nj by 4i(f) = (f(xj)),EJ. Then 4?
is an isomorphism of abelian groups (R-modules if R is commutative).
(4.11) Theorem. Let R be a commutative ring and let M and N be finitely

generated free R-modules. Then HomR(M, N) is a finitely generated free
R-module.
Proof. Let B = {v1, ... , v,, } be a basis of M and C = {w1, ... , wn} a basis
of N. Define f1 E HomR(M, N) for is 15 m and 1< j n by
_ fw, if k = i,
f',(vk) - 0 if k 76 i.
fij is a uniquely defined element of HomR(M, N) by Proposition 4.9.

We claim that { fij : 1 < i < m; 1 < j < n} is a basis of HomR(M, N).
To see this suppose that f E HomR(M, N) and for 1 < i < m write
f(vi) = ai1w1 + ... + ainwn.
Let
rm n
g = L, E aij fij.
i=1 j=1
Then
g(vk) = ak1wl + ... +aknwn = f(vk)
for 1 < k < m, so g = f since the two homomorphisms agree on a basis
of M. Thus, {fij 1 < i < m; I < j < n} generates HomR(M, N), and
:
we leave it as an exercise to check that this set is linearly independent and,

hence, a basis.
(4.12) Remarks.
(1) A second (essentially equivalent) way to see the same thing is to write
M ®;" 1R and N L" ®j=1R. Then, Corollary 3.13 shows that
m n
HomR(M, N) = ®®HomR(R, R).
i=1 j=1
But any f E HomR(R, R) can be written as f = f(1) . 1R. Thus
HomR(R, R) R so that
m n
HomR(M, N) L R.
i=1 j=1
(2) The hypothesis of finite generation of M and N is crucial for the va-
lidity of Theorem 4.11. For example, if R = Z and M = ®i°Z is the
free Z-module on the index set N, then Corollary 4.10 shows that
00
HomR(M, Z) L fl Z.
But the Z-module r IT Z is not a free Z-module. (For a proof of this fact
(which uses cardinality arguments), see I. Kaplansky, Infinite Abelian
Groups, University of Michigan Press, (1968) p. 48.)
(4.13) Proposition. Let M be a free R-module with basis S = {xj}jEJ. If

I is an ideal of R, then IM is a submodule of M and the quotient module
M/IM is an R/I-module. Let it : M --+ M/IM be the projection map.
Then M/IM is a free R/I-module with basis tr(S) = {Tr(xj)}jEJ.
Proof. Exercise. 0
(4.14) Proposition. Every R-module M is the quotient of a free module and

if M is finitely generated, then M is the quotient of a finitely generated free
R-module. In fact, we may take µ(F) = u(M).
Proof. Let S = {xj}jEJ be a generating set for the R-module M and let
F = ®jEJRj where Rj = R be the free R-module on the index set J. Define
the homomorphism TG : F -+ M by
,P((aj)jEJ) = Eajx3.
jEJ
Since S is a generating set for M, 0 is surjective and hence M LY F/ Ker(tp).

Note that if ISI < oo then F is finitely generated. (Note that every module
has a generating set S since we may take S = M.) Since M is a quotient of
F, we have µ(M) < µ(F). But F is free on the index set J (Remark 4.5),
so µ(F) < IJI, and since J indexes a generating set of M, it follows that
µ(F) < µ(M) if S is a minimal generating set of M. Hence we may take F
with µ(F) = µ(M). 0
(4.15) Definition. If M is an R-module then a short exact sequence

0---+ O
where F is a free R-module is called a free presentation of M.
Thus, Proposition 4.14 states that every module has a free presenta-
tion.
(4.16) Proposition. If F is a free R-module then every short exact sequence
0-+Ml -+M-f+F--i0
of R-modules is split exact.
Proof. Let S = {xj }jEJ be a basis of the free module F. Since f is surjective,
for each j E J there is an element yj E M such that f (yj) = x j . Define
h:S M by h(xj) = yj. By Proposition 4.9, there is a unique 6 E
HomR(F, M) such that f3ls = h. Since f o Q(xl) = xj = 1F(x,) for all

j E J, it follows that f o O = IF, and the result follows from Theorem
3.9. 0
(4.17) Corollary.
(1) Let M be an R-module and N C M a submodule with M/N free. Then
MAN®(M/N).
(2) If M is an R-module and F is a free R-module, then ML, Ker(f) ® F
for every surjective homomorphism f : M - F.
Proof. (1) Since M/N is free, the short exact sequence
0 --y N -- ' M M/N ---. 0
is split exact by Proposition 4.16 Therefore, M N ® (M/N) by Theorem
3.9.
(2) Take N = Ker(f) in part (1).
(4.18) Corollary. Let N be an arbitrary R-module and F a free R-module.

If
(4.1)
is a short exact sequence of R-modules, then
0 -. HomR(N, Ml) - HomR(N, M) -V1. HomR(N, F) -- 0

is a (split) short exact sequence of abelian groups (R-modules if R is com-
mutative).
Proof. By Proposition 4.16, the sequence (4.1) is split exact, so the corollary
follows immediately from Theorem 3.12.
(4.19) Remark. It is a theorem that any two bases of a free module over
a commutative ring R have the same cardinality. This result is proved
for finite-dimensional vector spaces by showing that any set of vectors of
cardinality larger than that of a basis must be linearly dependent. The
same procedure works for free modules over any commutative ring R, but
it does require the theory of solvability of homogeneous linear equations
over a commutative ring. However, the result can be proved for R a PID
without the theory of solvability of homogeneous linear equations over R;
we prove this result in Section 3.6. The result for general commutative rings
then follows by an application of Proposition 4.13.
The question of existence of a basis of a module, that is, to ask if a

given module is free, is a delicate question for a general commutative ring R.
We have seen examples of Z-modules, namely, finite abelian groups, which
are not free. We will conclude this section with the fact that all modules
over division rings, in particular, vector spaces, are free modules. In Section
3.6 we will study in detail the theory of free modules over a PID.
(4.20) Theorem. Let D be a division ring and let V be a D-module. Then

V is a free D-module. In particular, every vector space V has a basis.
Proof. The proof is an application of Zorn's lemma.
Let S be a generating set for V and let Bo C S be any linearly in-
dependent subset of S (we allow Bo = 0). Let T be the set of all linearly
independent subsets of S containing Bo and partially order T by inclusion.
If {Bi} is a chain in T, then UBi is a linearly independent subset of S
that contains Bo; thus, every chain in T has an upper bound. By Zorn's
lemma, there is a maximal element in T, so let B be a maximal linearly
independent subset of S containing Bo. We claim that S C (B) so that
V = (S) C (B). Let v E S. Then the maximality of B implies that V U {v}
is linearly dependent so that there is an equation
m
Eaivi+bv=0
i=i
where v1, ... , v,,, are distinct elements of B and al, ... , am, b E D are not
all 0. If b = 0 it would follow that Ei"_1 aivi = 0 with not all the scalars
ai = 0. But this contradicts the linear independence of B. Therefore, b 0 0
and we conclude
m
v = b-'(bv) = E(-b-lai)vi E (B).
i=i
Therefore, S C (B), and as observed above, this implies that B is a basis
of V.
The proof of Theorem 4.20 actually proved more than the existence of
a basis of V. Specifically, the following more precise result was proved.
(4.21) Theorem. Let D be a division ring and let V be a D-module. If S

spans V and Bo C S is a linearly independent subset, then there is a basis
BofVsuch that BOCBCS.
Proof.
(4.22) Corollary. Let D be a division ring, and let V be a D-module.

(1) Any linearly independent subset of V can be extended to a basis of V.
(2) A maximal linearly independent subset of V is a basis.
(3) A minimal generating set of V is a basis.
Proof. Exercise.
Notice that the above proof used the existence of inverses in the division
ring D in a crucial way. We will return in Section 3.6 to study criteria that
ensure that a module is free if the ring R is assumed to be a PID. Even
when R is a PID, e.g., R = Z, we have seen examples of R modules that
are not free, so we will still be required to put restrictions on the module
M to ensure that it is free.
3.5 Projective Modules

The property of free modules given in Proposition 4.16 is a very useful one,
and it is worth investigating the class of those modules that satisfy this
condition. Such modules are characterized in the following theorem.
(5.1) Theorem. The following conditions on an R-module P are equivalent.

(1) Every short exact sequence of R-modules
0- M, -+M-'P-.0
splits.
(2) There is an R-module P' such that P ® P is a free R-module.
(3) For any R-module N and any surjective R-module homomorphism ?p :
M - P, the homomorphism
tb.: HomR(N, M) -' HomR(N, P)
is surjective.
(4) For any surjective R-module homomorphism ¢ : M N, the homo-
morphism
0.: HomR(P, M) HomR(P, N)
is surjective.
Proof. (1) = (2). Let 0 -i K ---* F ---I P - 0 be a free presentation of

P. Then this short exact sequence splits so that F P ® K by Theorem
3.9.
(2) . (3). Suppose that F = P ® P' is free. Given a surjective R-
module homomorphism t : M P, let ?i' = V) ®1 P' : M ® P P ® P' _
F; this is also a surjective homomorphism, so there is an exact sequence
0--+Ker(V"') -i M®P' - F-+0.

Since F is free, Proposition 4.16 implies that this sequence is split exact;
Theorem 3.12 then shows that
?i; : HomR(N, M ® P) HomR(N, P 9 P')
3.5 Projective Modules 137
is a surjective homomorphism. Now let f E HomR(N, P) be arbitrary and

let f = i o f , where c : P -+ P ® P is the inclusion map. Then there
is an f E HomR(N, M ® P') with 10' (f) = f'. Let 7r : M ® P M
and 7r' : P ® P' P be the projection maps. Note that 7r' o c = lp and
0 o 7r = r' o V. Then
(7ro f) =V1o(?ro f)
=7r'o,'o f
=7r'0 f'
= (7rI oi.) o f
= f.
Therefore, j'. is surjective.
(3) (4). Let 0 --- K -. F "' P -+ 0 be a free presentation of
P. By property (3), there is a g E HomR(P, F) such that t',(Q) = 1p, i.e.,
loof3 = lp. Let 46: M -+ N be any surjective R-module homomorphism and
let f E HomR(P, N). Then there is a commutative diagram of R-module
homomorphisms
F - P -+ 0
If
M --' N -+ 0
with exact rows. Let S = {x,}jEJ be a basis of F. Since 0 is surjective,
we may choose yj E M such that 0(yj) = f o v/i(xj) for all j E J. By
Proposition 4.9, there is an R-module homomorphism g : F -+ M such
that g(xi) = y, for all j E J. Since 0 o g(x,) = 0(yy) = f o v/1(x, ), it follows
that D o g f o 0. Define f E HomR(P, M) by f= g -,3 and observe that
0.(f)=0o(go0)
=fo'YOQ
=folF
= f.
Hence, O.: HomR(P, M) - HomR(P, N) is surjective.
(4) => (1). A short exact sequence
0 - MI M -'p s P -+ 0,
in particular, includes a surjection V : M -i P. Now take N = P in part
(4). Thus,
ii.: HomR(P, M) -+ HomR(P, P)
is surjective. Choose /3 : P -+ M with t/i.(,0) = lp. Then /3 splits the short
exact sequence and the result is proved. 0
(5.2) Definition. An R-module P satisfying any of the equivalent conditions

of Theorem 5.1 is called projective.
As noted before Theorem 5.1, projective modules are introduced as the

class of modules possessing the property that free modules were shown to
possess in Proposition 4.16. Therefore, we have the following fact:
(5.3) Proposition. Free R-modules are projective.

Proof.
(5.4) Corollary. Let R be an integral domain. If P is a projective R-module,

then P is torsion-free.
Proof. By Theorem 5.1 (2), P is a submodule of a free module F over R.
According to Proposition 4.8, every free module over an integral domain is
torsion-free, and every submodule of a torsion-free module is torsion-free.
(5.5) Corollary. An R-module P is a finitely generated projective R-module

if and only if P is a direct summand of a finitely generated free R-module.
Proof. Suppose that P is finitely generated and projective. By Proposition
4.14, there is a free presentation
0- K - F , P - 0
such that F is free and p(F) = µ(P) < oo. By Theorem 5.1, P is a direct
summand of F.
Conversely, assume that P is a direct summand of a finitely generated
free R-module F. Then P is projective, and moreover, if P ® P - F then
F/P - P so that P is finitely generated.
(5.6) Examples.
(1) Every free module is projective.
(2) Suppose that m and n are relatively prime natural numbers. Then
as abelian groups Z,,,,, ?5 Z,,, ® Z,,. It is easy to check that this iso-
morphism is also an isomorphism of Z,,,,,-modules. Therefore, Z,,, is
a direct summand of a free Z,,,,,-module, and hence it is a projective
Z,,,,,-module. However, Z,,, is not a free Z,,,,, module since it has fewer
than mn elements.
(3) Example 5.6 (2) shows that projective modules need not be free. We
will present another example of this phenomenon in which the ring R is
an integral domain so that simple cardinality arguments do not suffice.
Let R = Z(v5] and let I be the ideal I = (2, 1 + = (al, a2). It
is easily shown that I is not a principal ideal, and hence by Example
4.6 (6), we see that I cannot be free as an R-module. We claim that I
is a projective R-module. To see this, let b = 1 - E R, let F be a

free R-module with basis Is,, 82}, and let 0: F - I be the R-module
homomorphism defined by
c(rjs1 + r2s2) = r1al + r2a2.
Now define an R-module homomorphism a : 1 -+ F by

a(a) = -as, + ((ab)/2)s2.
Note that this makes sense because 2 divides ab for every a E I. Now
for a E I,
0 o a(a) = b(-asl + ((ab)/2)s2)
= -aal + ((ab)/2)a2
= -aal + aa2b/2
= -2a + 3a
=a
so that a is a splitting of the surjective map 0. Hence, F Ker(O) ® I
and by Theorem 5.1, I is a projective R-module.
Concerning the construction of new projective modules from old ones,

there are the following two simple facts:
(5.7) Proposition. Let { Pj }jEJ be a family of R-modules, and let P =

®jEJ Pi. Then P is projective if and only if Pj is projective for each j E J.
Proof. Suppose that P is projective. Then by Theorem 5.1, there is an
R-module P' such that P ® P = F is a free R-module. Then
F=P®P'=(®Pj)ED P',
jEJ
and hence, each Pj is also a direct summand of the free R-module F. Thus,
Pj is projective.
Conversely, suppose that Pj is projective for every j E J and let P,' be
an R-module such that Pj ® Pj' = F, is free. Then
P®((DPj')(P,®P;)
jEJ jEJ
® Fj.
jEJ
Since the direct sum of free modules is free (Example 4.6 (8)), it follows
that P is a direct summand of a free module, and hence P is projective.
(5.8) Proposition. Let R be a commutative ring and let P and Q be finitely

generated projective R-modules. Then HomR(P, Q) is a finitely generated
projective R-module.
Proof. Since P and Q are finitely generated projective R-modules, there are
R-modules P and Q' such that P®P' and Q®Q' are finitely generated free
modules. Therefore, by Theorem 4.11, HomR(P ® P, Q (D Q') is a finitely
generated free R-module. But
HomR(P (D P', Q ® Q') ?' HomR(P, Q) ® HomR(P, Q')
® HomR(P', Q) ® HomR(P, Q')
so that HomR(P, Q) is a direct summand of a finitely generated free R-
module, and therefore, it is projective and finitely generated by Corollary
5.5. 0
Example 5.6 (3) was an example of an ideal in a ring R that was
projective as an R-module, but not free. According to Example 4.6 (6), an
ideal I in a ring R is free as an R-module if and only if the ideal is principal.
It is a natural question to ask which ideals in a ring R are projective as
R-modules. Since this turns out to be an important question in number
theory, we will conclude our brief introduction to the theory of projective
modules by answering this question for integral domains R.
(5.9) Definition. Let R be an integral domain and let K be the quotient

field of R. An ideal I C R is said to be invertible if there are elements
a1, ... , an E I and bl, ... , bn E K such that
(5.1) b=I C_ R for 1 < i < n, and
(5.2) albs + + anbn = 1.
(5.10) Examples.
(1) If I C R is the principal ideal I = (a) where a 0 0, then I is an
invertible ideal. Indeed, let b = 1/a E K. Then any x E I is divisible
by a in R so that bx = (1/a)x E R, while a(l/a) = 1.
(2) Let R = Z[vr---5] and let I = (2, 1 + ). Then it is easily checked
that I is not principal, but I is an invertible ideal. To see this, let
a1 =2,a2=1+-,,/--5,b1 =-1, and b2 = (1 - -,/--5)/2. Then
albs + a2b2 = -2 + 3 = 1.
Furthermore, a1b2 and a2b2 are in R, so it follows that b2I C R, and
we conclude that I is an invertible ideal.
The following result characterizes which ideals in an integral domain

R are projective modules. Note that the theorem is a generalization of
Example 5.6 (3):
(5.11) Theorem. Let R be an integral domain and let I C R be an ideal.

Then I is a projective R-module if and only if I is an invertible ideal.
Proof. Suppose that 1 is invertible and choose al, ... , an E I and b1, ... , bn
in the quotient field K of R so that Equations (5.1) and (5.2) are satisfied.
Let 0: Rn -' I be defined by
O(xi, ... , xn) = alxl + ... + anxn,
and define 6 : I -+ Rn by
/3(a) = (abl, ... , abn).
Note that abi E R for all i by Equation (5.1). Equation (5.2) shows that
n
O o /3(a) = E ai(abi) = a t aibi =a
i_1 i=1
for every a E I. Therefore 0 o /j = 1 p and Theorem 3.9 implies that I is a

direct summand of the free R-module Rn, so I is a projective R-module.
Conversely, assume that the ideal 19 R is projective as an R-module.
Then I is a direct summand of a free R-module F, so there are R-module
homomorphisms 0 : F -+ I and 3 : I -' F such that 0 o /3 = 11. Let
S = {xj}jEJ be a basis of F. Given x E 1, /3(x) E F can be written
uniquely as
(5.3) /3(x) _ cjxj.

jEJ
For each j E J, let tlj(x) = cj. This gives a function Oj : I R, which is
easily checked to be an R-module homomorphism. If aj = O(xj) E I, note
that
(5.4) for each x E I, Oj(x) = 0 except for at most finitely many j E J;
(5.5) for each x E I, Equation (5.3) shows that
x = -0(0(x)) = 1: 'pj(x)aj
jEJ
Given x j4 0 E I and j E J, define b, E K (K is the quotient field of

R) by
(5.6) bj = Oj(x)
X
The element bj E K depends on j E J but not on the element x 54 0 E I.
To see this, suppose that x' 54 0 E I is another element of I. Then
X'pj(x) = Oj(x'x) =''j(xx) = x1pj(x )
so that >' 1(x)/x = 1(ii(x')/x'. Therefore, for each j E J we get a uniquely

defined bi E K. By property (5.4), at most finitely many of the bi are not
0. Label the nonzero bi by b1, ... , bn. By property (5.5), if x $ 0 E I then
n n n
x= E V)i(x)a.i =>(bix)ai =x (bia)).
.i=1 .i=1 i=1
Cancelling x $ 0 from this equation gives

a1b1 + ... + anbn = 1
where a1, ... , an E I and b1 , bn E K. It remains to check that b,I C R

for 1 < j < n. But if x $ 0 E I then bi = tki(x)/x so that bix = V)j(x) E R.
Therefore, I is an invertible ideal and the theorem is proved. 0
(5.12) Remark. Integral domains in which every ideal is invertible are known
as Dedekind domains, and they are important in number theory. For ex-
ample, the ring of integers in any algebraic number field is a Dedekind
domain.
3.6 Free Modules over a PID

In this section we will continue the study of free modules started in Sec-
tion 3.4, with special emphasis upon theorems relating to conditions which
ensure that a module over a PID R is free. As examples of the types of
theorems to be considered, we will prove that all submodules of a free R-
module are free and all finitely generated torsion-free R-modules are free,
provided that the ring R is a PID. Both of these results are false without
the assumption that R is a PID, as one can see very easily by consider-
ing an integral domain R that is not a PID, e.g., R = Z[X], and an ideal
I C R that is not principal, e.g., (2, X) C Z[XJ. Then I is a torsion-free
submodule of R that is not free (see Example 4.6 (6)).
Our analysis of free modules over PIDs will also include an analysis of
which elements in a free module M can be included in a basis and a criterion
for when a linearly independent subset can be included in a basis. Again,
these are basic results in the theory of finite-dimensional vector spaces, but
the case of free modules over a PID provides extra subtleties that must be
carefully analyzed.
We will conclude our treatment of free modules over PIDs with a fun-
damental result known as the invariant factor theorem for finite rank sub-
modules of free modules over a PID R. This result is a far-reaching gener-
alization of the freeness of submodules of free modules, and it is the basis
3.6 Free Modules over a PID 143
for the fundamental structure theorem for finitely generated modules over
PIDs which will be developed in Section 3.7.
We start with the following definition:
(6.1) Definition. Let M be a free R-module. Then the free rank of M,

denoted free-rankR(M), is the minimal cardinality of a basis of M.
Since we will not be concerned with the fine points of cardinal arith-
metic, we shall not distinguish among infinite cardinals so that
free-rankR(M) E Z+ U {oo}.
Since a basis is a generating set of M, we have the inequality u(M) <
free-rankR(M). We will see in Corollary 6.18 that for an arbitrary commu-
tative ring R and for every free R-module, free-rankR(M) = p(M) and all
bases of M have this cardinality.
(6.2) Theorem. Let R be a PID, and let M be a free R-module. If N C M

is a submodule, then N is a free R-module, and
free-rankR(N) < free-rankR(M).
Proof. We will first present a proof for the case where free-rankR(M) < 00.
This case will then be used in the proof of the general case. For those who
are only interested in the case of finitely generated modules, the proof of
the second case can be safely omitted.
Case 1. free-rankR(M) < 00.
We will argue by induction on k = free-rankR(M). If k = 0 then
M = (0) so N = (0) is free of free-rank 0. If k = 1, then M is cyclic so
M = (x) for some nonzero x E M. If N = (0) we are done. Otherwise, let
I = {a ER: ax EN). Since! is an ideal ofRandRisaPID,I=(d);
since N j4 (0), d 54 0. If Y E N then y = ax = rdx E (dx) so that N = (dx)
is a free cyclic R-module. Thus free-rankR(N) = 1 and the result is true
for k = 1.
Assume by induction that the result is true for all M with free-rank k,
and let M be a module with free-rankR(M) = k+l. Let S = {x1, ... xk+l }
be a basis of M and let Mk = (xl, ... ,xk). If N C Mk we are done by
induction. Otherwise N n Mk is a submodule of Mk which, by induction, is
free of free-rank e < k. Let {yl, ... , ye} be a basis of N n Mk. By Theorem
2.5
N/(N n Mk) 25 (N + Mk)/Mk C M/Mk = (xk+I + Mk)-
By the k = 1 case of the theorem, (N + Mk)/Mk is a free cyclic submodule
of M/Mk with basis dxk+1 + Mk where d 54 0. Choose ye+l E N so that
ye+i = dxk+i + x' for some x' E Mk. Then (N + Mk) lMk = (ye+i + Mk).
We claim that S' = {yl, ... , ye, yt+1 } is a basis of N. To see this, let y E N.
Then y + Mk = ae+i(ye+i + Mk) so that y - at+lyt+1 E N n Mk, which
implies that y - at+Iyt+1 = a1y1 + aeye. Thus S' generates N. Suppose
that aly, + +ae+lyt+1 = 0. Then at+1(dxk+1 +x')+aIy1 + +aeye = 0
so that ae+ldxk+l E Mk. But S is a basis of M so we must have ae+ld = 0;
since d 34 0 this forces at+1 = 0. Thus a1y1 + + aeye = 0 which implies
that a1 = = at = 0 since {y1, . . , yt} is linearly independent. Therefore
S' is linearly independent and hence a basis of N, so that N is free with
free-rankR(N) < e + 1 < k + I. This proves the theorem in Case 1.
Case 2. free-rankn(M) = oo.
Since (0) is free with basis 0, we may assume that N # (0). Let S =
{xj}jEJ be a basis of M. For any subset K C J let MK = ({xk}kEK)
and let NK = N n MK. Let T be the set of all triples (K, K', f) where
K' C K C J and f : K' -+ NK is a function such that (f (k)}kEK' is a
basis of NK We claim that T # 0.
.
Since N $ (0) there is an x j4 0 E N, so we may write x = aixj, + +

akxjk. Hence X E NK where K = {{ j 1 ,---, }. But MK is a free R-module
with free-rankR(MK) < k < oo and NK is a nonzero submodule. By Case
1, NK is free with free-rankR(NK) = e < k. Let {y1, ... yt} be a basis of
NK, and let K' = (j 1, ... , jt}, and define f : K' NK by f (j;) = y, for
1 < i < e. Then (K, K', f) E T so that T A 0, as claimed.
Now define a partial order on T by setting (K, K', f) < (L, L', g) if
K C L, K' C L', and 9IK' = f. If {(KQ, KQ, fQ)}QEA C T is a chain, then
(UQEA KQ, UQEA KQ, F) where FIK' = fQ is an upper bound in T for
the chain. Therefore, Zorn's lemma applies and there is a maximal element
(K, K', f) of T.
Claim. K = J.
Assuming the claim is true, it follows that MK = M, NK = N n MK =
N, and {f(k)}kEK' is a basis of N. Thus, N is a free module (since it has
a basis), and since S was an arbitrary basis of M, we conclude that N has
a basis of cardinality < free-rankR(M), which is what we wished to prove.
It remains to verify the claim. Suppose that K j4 J and choose j E
J \ K. Let L = K U {j}. If NK = NL then (K, K', f) (L, K', f),
contradicting the maximality of (K, K', f) in T. If NK 36 NL, then
NL/(NL n MK) (NL + MK)/MK C ML/MK = (x3 + MK)-
By Case 1, (NL + 1MIK)/MK is a free cyclic submodule with basis dxj + MK
where d # 0. Choose z E NL so that z = dxj + w for some w E MK.
Then (NL + MK)/MK = (z + MK). Now let L' = K' U {j} and define
L' -+NL by
f,(k) = J f (k) if k E K',
Z ifk=j.
We need to show that { f'(k)}kEL' is a basis of NL. But if x E NL then

x + MK = cz + MK for some c E R. Thus x - cz E MK n N = NK so that
x - cz = bkf(k)
kE K'
where bk E R. Therefore, { f (k)}kEL' generates NL.

Now suppose EkEL' bkf'(k) = 0. Then
b,z+1: bkf(k)=0
kE K'
so that
db,x,+b,w+ E bkf(k) =0.
kEK'
That is, db,x, E MK n (x,) = (0), and since S = {xt}tEJ is a basis
of M, we must have db, = 0. But d j4 0, so b, = 0. This implies that
EkE K, bk f (k) = 0. But { f (k) }kE K' is a basis of NK, so we must have
bk = 0 for all k E K'. Thus { f'(k)}kEL' is a basis of NL. We conclude that
(K, K', f) (L, L', f'), which contradicts the maximality of (K, K', f).
Therefore, the claim is verified, and the proof of the theorem is complete.
(6.3) Corollary. Let R be a PID and let P be a projective R-module. Then

P is free.
Proof. By Proposition 4.14, P has a free presentation
0--+K-+F--+P-+0.
Since P is projective, this exact sequence splits and hence F P ® K.
Therefore, P is isomorphic to a submodule of F, and Theorem 6.2 then
shows that P is free.
(6.4) Corollary. Let M be a finitely generated module over the PID R and
let N C M be a submodule. Then N is finitely generated and
µ(N) < µ(M).
Proof. Let
0- K- F 0+M-y0
be a free presentation of M such that free-rank(F) = p(M) < oo, and let
NI = 0-1(N). By Theorem 6.2, N1 is free with
µ(N1) < free-rank(NI) < free-rank(F) = µ(M).
Since N = O(NI ), we have µ(N) < µ(N1), and the result is proved.
(6.5) Remark. The hypothesis that R be a PID in Theorem 6.2 and Corol-
laries 6.3 and 6.4 is crucial. For example, consider the ring R = Z[X] and
let M = R and N = (2, X). Then M is a free R-module and N is a sub-
module of M that is not free (Example 4.6 (6)). Moreover, R = Z[V/-5],
P = (2, 1 + /) gives an example of a projective R-module P that is
not free (Example 5.6 (3)). Also note that 2 = µ(N) > µ(M) = 1 and
2 = µ(P) > I = u(R).
Recall that if M is a free module over an integral domain R, then M is
torsion-free (Proposition 4.8). The converse of this statement is false even
under the restriction that R be a PID. As an example, consider the Z-
module Q. It is clear that Q is a torsion-free Z-module, and it is a simple
exercise to show that it is not free. There is, however, a converse if the
module is assumed to be finitely generated (and the ring R is a PID).
(6.6) Theorem. If R is a PID and M is a finitely generated torsion-free

R-module, then M is free and
free-rankR(M) = µ(M).
Proof. The proof is by induction on µ(M). If µ(M) = 1 then M is cyclic

with generator {x}. Since M is torsion-free, Ann(x) = {0}, so the set {x}
is linearly independent and, hence, is a basis of M.
Now suppose that µ(M) = k > 0 and assume that the result is true
for all finitely generated torsion-free R-modules M' with µ(M') < k. Let
{x1, ... , xk } be a minimal generating set for M, and let
M1={xERI:axE(x1) for some a0OER}.

Then M/M1 is generated by {x2 + M1, ... , xk + M1 } so that µ(M/Ml) _
j < k - 1. If ax E M1 for some a 54 0 E R, then from the definition of M1,
b(ax) E (xl) for some b 0 0. Hence x E M1 and we conclude that M/Ml
is torsion-free. By the induction hypothesis, M/M1 is free of free-rank j.
Then Corollary 4.17 shows that ML, M1® (M/Ml ). We will show that M1
is free of free-rank 1. It will then follow that
k=µ(l41)<p(Mi)+µ(MIMI)=1+j,
and since j < k - 1, it will follow that j = k -1 and M is free of free-rank =
k.
It remains to show that M1 is free of rank 1. Note that if R is a field
then M1 = R xl and we are done. In the general case, M1 is a submodule
of M, so it is finitely generated by t < k elements. Let {y', ... , yt} be
a generating set for M1 and suppose that a;y; = bix1 with a, # 0 for
1 < i < t. Let qo = a1 ... at.
Claim. If ax = bxl with a 34 0 then a I bqo
To see this note that x = Ei=1 ciyi so that
qox = cjgoy;
;=1
_ ci(go/ai)aiyi
i=1
_ G(qo/a;)b;xi
;=1
I
Ec;(go/a;)b; x1
i=1
Therefore.
I
bgox, = agox = (c(o/ai)bi)
aE X1.
i=1
Since M1 is torsion-free, it follows that

t
bqo = a (ci(o/a)bi)
i=1
and the claim is proved.

Using this claim we can define a function : M1 --- R by t(x) _
(bqo)/a whenever ax = bx1 for a 36 0. We must show that 0 is well defined.
That is, if ax = bx1 and a'x = b'x, then (bqo)/a = (b'qo)/a'. But ax = bx1
and a'x = b'x1 implies that a'bx1 = a'ax = ab'x1 so that alb = all because
M is torsion-free. Thus a'bqo = ab'qo so that (bqo)/a = (b'qo)/a' and 0 is
well defined. Furthermore, it is easy to see that 0 is an R-module homo-
morphism so that Im(O) is an R-submodule of R, i.e., an ideal. Suppose
that O(x) = 0. Then ax = bx1 with a 0 0 and O(x) = (bqo)/a = 0 E R.
Since R is an integral domain, it follows that b = 0 and hence ax = 0. Since
M is torsion-free we conclude that x = 0. Therefore, Ker(O) = {0} and
M1 = Im(O) = Rc.
Hence, M1 is free of rank 1, and the proof is complete.
(6.7) Corollary. If M is a finitely generated module over a field F, then M

is free.
Proof. Every module over a field is torsion-free (Proposition 2.20).
(6.8) Remark. We have already given an independent proof (based on Zorn's

lemma) for Corollary 6.7, even without the finitely generated assumption
(Theorem 4.20). We have included Corollary 6.7 here as an observation that
it follows as a special case of the general theory developed for torsion-free

finitely generated modules over a PID.
(6.9) Corollary. If M is a finitely generated module over a PID R, then

(M/M,).
Proof. There is an exact sequence of R-modules
0-+M,--iM-'M/M, 0.
Hence, M/M, is finitely generated and by Proposition 2.18, it is torsion-

free, so Theorem 6.6 shows that M/M, is free. Then Corollary 4.17 shows
that M M, ® (M/M,).
The main point of Corollary 6.9 is that any finitely generated module
over a PID can be written as a direct sum of its torsion submodule and
a free submodule. Thus an analysis of these modules is reduced to study-
ing the torsion submodule, once we have completed our analysis of free
modules. We will now continue the analysis of free modules over a PID R
by studying when an element in a free module can be included in a basis.
As a corollary of this result we will be able to show that any two bases
of a finitely generated free R-module (R a PID) have the same number of
elements.
(6.10) Example. Let R be a PID and view R as an R-module. Then an

element a E R forms a basis of R if and only if a is a unit. Thus if R is
a field, then every nonzero element is a basis of the R-module R, while if
R = Z then the only elements of Z that form a basis of Z are 1 and -1.
As a somewhat more substantial example, consider the Z-module Z2. Then
the element u = (2, 0) E Z2 cannot be extended to a basis of Z2 since if v
is any element of Z2 with {u, v} linearly independent, the equation
au+f3v=(1,0)
is easily seen to have no solution a, 0 E Z. Therefore, some restriction on
elements of an R-module that can be included in a basis is necessary. The
above examples suggest the following definition.
(6.11) Definition. Let M be an R-module. A torsion-free element x # 0 E M

is said to be primitive if x = ay for some y E M and a E R implies that a
is a unit of R.
(6.12) Remarks.
(1) If R is a field, then every nonzero x E M is primitive.
(2) The element x E R is a primitive element of the R-module R if and
only if x is a unit.
(3) The element (2, 0) E Z2 is not primitive since (2, 0) = 2 (1, 0).
(4) If R = Z and M = Q, then no element of M is primitive.
(6.15) Lemma. Let R be a PID and let M be a free R-module with basis
S = {xj}jEJ. If X = E,EJajxj E M, then x is primitive if and only if
gcd(jai }jEJ) = 1.
Proof. Let d = gcd({aj}jEJ). Then x = d(>2jEJ(aj/d)xj), so if d is not a
unit then x is not primitive. Conversely, if d = 1 and x = ay then
E ajxj = x
jEJ
= ay(
=
l
a(Ebjxj)
jEJ
_ E abjxj.
jEJ
Since S = {xj}jEJ is a basis, it follows that aj = abj for all j E J. That

is, a is a common divisor of the set {aj}jEJ so that a I d = 1. Hence a is a
unit and x is primitive.
(6.14) Lemma. Let R be a PID and let M be a finitely generated R-module.

If x E M has Ann(x) = (0), then we may write x = ax' where a E R and
x' is primitive. (In particular, if M is not a torsion module, then M has a
primitive element.)
Proof. Let xo = x. If xo is primitive we are done. Otherwise, write xo = alxl
where al E R is not a unit. Then (xo) 5 (x1). To see this, it is certainly
true that (xo) C (xi). If the two submodules are equal then we may write
xl = bxo so that xo = aixl = albxo, i.e., (1 - alb) E Ann(xo) = (0).
Therefore, 1 = alb and al is a unit, which contradicts the choice of al.
Now consider xl. If xl is primitive, we are done. Otherwise xl =
a2x2 where a2 is not a unit, and as above we conclude that (xl) (x2).
Continuing in this way we obtain a chain of submodules
(6.1) (xo) 5 (xi) 5 (x2) 5 ...
Either this chain stops at some i, which means that xi is primitive, or (6.1)
is an infinite properly ascending chain of submodules of M. We claim that
the latter possibility cannot occur. To see this, let N = Ui_° I (xi). Then N
is a submodule of the finitely generated module M over the PID R so that
N is also finitely generated by {yl,... , yk } (Corollary 6.4). Since (xo) C
(xl) C , there is an i such that {yi,...,yk} C (xi). Thus N = (xi) and
hence (x,) = (xi+i) = , which contradicts having an infinite properly
ascending chain. Therefore, xi is primitive for some i, and if we let x' = xi

we conclude that x = ax' where a = ala2 . ai. 0
(6.15) Remark. Suppose that M is a free R-module, where R is a PID, and
x E M. Then Ann(x) = (0), so x = ax' where x' is a primitive element of
M. If S = {xj }jEJ is a basis of M, then we may write x' = E,E J bjxj so
that
x = ax=Eabjxj= cjxj.
jEJ jEJ
Since gcd({bj}jEJ) = 1 (by Lemma 6.13) we see that a = gcd({cj}jEJ).
The element a E R, which is uniquely determined by x up to multiplication
by a unit of R, is called the content of x E M and is denoted c(x). (Compare
with the concept of content of polynomials (Definition 2.6.3).) Thus, any
x E M can be written
(6.2) x = c(x) i
where x' is primitive.
(6.16) Theorem. Let R be a PID and let M be a fee R-module with

rank(M) = k = µ(M) = free-rank(M).
If x E M is primitive, then M has a basis of k elements containing x.
Proof. Assume first that k < oo and proceed by induction on k. Suppose
k = 1 and let M have a basis {x1 }. Then x = ax1 for some a E R. Since x
is primitive, it follows that a is a unit so that (x) = (x1) = M, hence {x}
is a basis of M.
The case k = 2 will be needed in the general induction step, so we
present it separately. Thus suppose that M has a basis {x1, x2} and let
x = rx1 + sx2 where r, s E R. Since x is primitive, gcd{r, s} = 1, so we
may write ru + sv = 1. Let x2 = -vxl + ux2. Then
x1 =ux-sx2
and
x2 = vx + rx2.
Hence, (x, x2) = M. It remains to show that {x, x2} is linearly indepen-
dent. Suppose that ax + bx2 = 0. Then
a(rx1 + sx2) + b(-vx1 + UX2) = 0.
Since {x1i x2} is a basis of M, it follows that
ar-bv=0
and
as+bu=0.
Multiplying the first equation by u, multiplying the second by v, and adding
shows that a = 0, while multiplying the first by -s, multiplying the second
by r, and adding shows that b = 0. Hence, {x, x2} is linearly independent
and, therefore, a basis of M.
Now suppose that u(M) = k > 2 and that the result is true for all free
R-modules of rank < k. By Theorem 6.6 there is a basis {x1, ... , xk } of M.
Letx=Ekla;x;.Ifak=0then xEM1=(x1,...,xk_1),sobyinduc-
tion there is a basis {x, x2, ... ,xk_1} of M1. Then {x,x2, ... ,xk_1, xk} is
a basis of M containing x. Now suppose that ak # 0 and let y = Ei=1 a;x;.
If y = 0 then x = akxk, and since x is primitive, it follows that ak is a unit
of R and {x1, ... ,xk_1, x} is a basis of M containing x in this case. If
y 0 then there is a primitive y' such that y = by' for some b E R. In
particular, y' E M1 so that M1 has a basis {y', x2, ... , x'ti-11 and hence
M has a basis {y', x2, ... , xk_1, xk}. But x =akxk + y = akxk + by' and
gcd(ak, b) = 1 since x is primitive. By the previous case (k = 2) we conclude
that the submodule (xk, y') has a basis {x, y"}. Therefore, M has a basis
{x, x2, ... , xk_1, y"} and the argument is complete when k = µ(M) < oo.
If k = oo let {x, }jE J be a basis of M and let x = 1 aixj, for
some finite subset I = 01, ... , j } c J. If N = (xj...... xi,,) then x is
a primitive element in the finitely generated module N, so the previous
argument applies to show that there is a basis {x, x2, ... , x' } of N. Then
{x, x2, ... ,xn} U {x,}iEJ\I is a basis of M containing x. 0
(6.17) Corollary. If M is a free module over a PID R, then every basis of
M contains µ(M) elements.
Proof. In case µ(M) < oo, the proof is by induction on µ(M). If µ(M) = 1
then M = (x). If {x1, x2} C M then x1 = a1x and and x2 = a2x so that
a2x1 -a, X2 = 0, and we conclude that no subset of M with more than one
element is linearly independent.
Now suppose that p(M) = k > 1 and assume the result is true for all
free R-modules N with µ(N) < k. Let S = {x) }JE J c M be any basis of
M and choose x E S. Since x is primitive (being an element of a basis),
Theorem 6.16 applies to give a basis {x, y2i ... , yk } of M with precisely
µ(M) = k elements. Let N = M/(x) and let it : M - N be the projection
map. It is clear that N is a free R-module with basis 7r(S) \ {ar(x)}. By
Proposition 2.12 it follows that µ(N) > k -1, and since {7r(y2), ... , ir(yk)}
generates N, we conclude that µ(N) = k - 1. By induction, it follows that
ISI - 1 < oo and ISO - 1 = k - 1, i.e., ISI = k, and the proof is complete in
case µ(M) < oo.
In case µ(M) = oo, we are claiming that no basis of M can contain a
finite number k E Z+ of elements. This is proved by induction on k, the
proof being similar to the case µ(M) finite, which we have just done. We
leave the details to the reader. 0
(6.18) Corollary. Let R be any commutative ring with identity and let M be
a free R-module. Then every basis of M contains µ(M) elements.
Proof. Let I be any maximal ideal of R (recall that maximal ideals exist
by Theorem 2.2.16). Since R is commutative, the quotient ring R/I = K
is a field (Theorem 2.2.18), and hence it is a P1D. By Proposition 4.13,
the quotient module M/IM is a finitely generated free K-module so that
Corollary 6.17 applies to show that every basis of M/IM has µ(M/IM)
elements. Let S = {xj}JEJ be an arbitrary basis of the free R-module M
and let it : M - M/IM be the projection map. According to Proposition
4.13, the set 7r(S) = {7r(xJ)}3EJ is a basis of M/IM over K, and therefore,
14(m)< V1 = µ(M/IM) 5 AM.

Thus, µ(M) = IJI, and the corollary is proved. 0
(6.19) Remarks.
(1) If M is a free R-module over a commutative ring R, then we have
proved that free-rank(M) = µ(M) = the number of elements in any
basis of M. This common number we shall refer to simply as the rank
of M, denoted rankR(M) or rank(M) if the ring R is implicit. If R is
a field we shall sometimes write dimR(M) (the dimension of M over
R) in place of rankR(M). Thus, a vector space M (over R) is finite
dimensional if and only if dimR(M) = rankR(M) < oo.
(2) Corollary 6.18 is the invariance of rank theorem for finitely generated
free modules over an arbitrary commutative ring R. The invariance of
rank theorem is not valid for an arbitrary (possibly noncommutative)
ring R. As an example, consider the Z-module M = ®nENZ, which
is the direct sum of countably many copies of Z. It is simple to check
that M M ® M. Thus, if we define R = Endz(M), then R is a
noncommutative ring, and Corollary 3.13 shows that
R = Endz(M)
= Homz(M, M)
Homz(M, M (D M)
Homz(M, M) ® Homz(M, M)
R®R.
The isomorphisms are isomorphisms of Z-modules. We leave it as an
exercise to check that the isomorphisms are also isomorphisms of R-
modules, so that R a, R2, and hence, the invariance of rank does
not hold for the ring R. There is, however, one important class of
noncommutative rings for which the invariance of rank theorem holds,
namely, division rings. This will be proved in Proposition 7.1.14.
(6.20) Corollary. If M and N are free modules over a PID R, at least one of
which is finitely generated, then M N if and only if rank(M) -= rank(N).
Proof. If M and N are isomorphic, then p(M) = µ(N) so that rank(M) =
rank(N). Conversely, if rank(M) = rank(N), then Proposition 4.9 gives a
homomorphism f : M - N, which takes a basis of M to a basis of N. It is
easy to see that f must be an isomorphism.
(6.21) Remark. One of the standard results concerning bases of finite-

dimensional vector spaces is the statement that a subset S = {x1, ... , x")
of a vector space V of dimension n is a basis provided that S is either a
spanning set or linearly independent. Half of this result is valid in the cur-
rent context of finitely generated free modules over a PID. The set (2) C Z
is linearly independent, but it is not a basis of the rank 1 Z-module Z.
There is, however, the following result.
(6.22) Proposition. Let M be a finitely generated free R-module of rank = k

where R is a PID. If S = {x1, ... ,xk} generates M, then S is a basis.
Proof. Let T = {e1}?-1 be the standard basis of Rk. Then there is a homo-
morphism 0: Rk - M determined by 4,(e,) = xj. Since (S) = M, there is
a short exact sequence
RkM ---+ O
where K = Ker(¢). Since M is free, Corollary 4.16 gives Rk -- M ®K, and
according to Theorem 6.2, K is also free of finite rank. Therefore,
k = rank(M) + rank(K) = k + rank(K)

and we conclude that rank(K) = 0. Hence 0 is an isomorphism and S is a
basis.
We will conclude this section with a substantial generalization of The-

orem 6.2. This result is the crucial result needed for the structure theorem
for finitely generated modules over a PID.
(6.23) Theorem. (Invariant factor theorem for submodules) Let R be a

PID, let M be a free R-module, and let N C M be a submodule (which is
automatically free by Theorem 6.2) of rank n < oo. Then there is a basis
S of M, a subset {x1, ... , xn} C S, and nonzero elements ,, ..., sn E R
such that
(6.3) {s1x1, ... ,snx,a} is a basis of N

and
(6.4) si I si+1 for 1 < i < n - 1.
Proof. If N = (0), there is nothing to prove, so we may assume that N 96 (0)

and proceed by induction on n = rank(N). If n = 1, then N = (y) and {y}
is a basis of N. By Lemma 6.14, we may write y = c(y)x where x E M is a
primitive element and c(y) E R is the content of y. By Theorem 6.16, there
is a basis S of M containing the primitive element x. If we let x1 = x and
Si = c(y), then slxl = y is a basis of N, so condition (6.3) is satisfied; (6.4)
is vacuous for n = 1. Therefore, the theorem is proved for n = 1.
Now assume that n > 1. By Lemma 6.14, each y E N can be written as
y = c(y) y' where c(y) E R is the content of y (Remark 6.15) and y' E M
is primitive. Let
S = {(c(y)) : y E N}.
This is a nonempty collection of ideals of R. Since R is Noetherian, Propo-
sition 2.5.10 implies that there is a maximal element of S. Let (c(y)) be
such a maximal element. Thus, y E N and y = c(y) x, where x E M is
primitive. Let s1 = c(y). Choose any basis T of M that contains x. This is
possible by Theorem 6.16 since x E M is primitive. Let xl = x and write
T = {x1}UT' = {xl}U{x'j}jEJ'. Let M1 = ({x'j}jEJ') and let N1 = M1f1N.
Claim. N = (slxl) ®N1.
To see this, note that (six,) fl Nl C (x1) fl M1 = (0) because T is a

basis of M. Let z E N. Then, with respect to the basis T, we may write
z = alx1 + b
jEJ'
Let d = (Si, al) = gcd{s1, al}. Then we may write d = us1 +val where u,
V E R. If w = uy + vz, then Equation (6.5) shows that
w=uy+vz
= (us1 + val)xl + E vbjx'j
jEJ'
= dx1 + = vbjxj.
jEJ'
Writing w = c(w) w' where c(w) is the content of w and w' E M is
primitive, it follows from Lemma 6.13 that c(w) I d (because c(w) is the
greatest common divisor of all coefficients of w when expressed as a linear
combination of any basis of M). Thus we have a chain of ideals
(Si) C (d) C (c(w)),
and the maximality of (sl) in S shows that (Si) = (c(w)) = (d). In partic-
ular, (Si) = (d) so that s1 I a1, and we conclude that
z = b1(slxl) + E bjij
jEJ'
That is, z E (s1x1) + N1. Theorem 3.1 then shows that
N = (six,) ® Ni,
and the claim is proved.
By Theorem 6.2, N1 is a free R-module since it is a submodule of the
free R-module M. Furthermore, by the claim we see that
rank(N1) = rank(N) - 1 = n - 1.
Applying the induction hypothesis to the pair N1 C M1, we conclude that
there is a basis S' of M1 and a subset {x2, ... xn} of S', together with
nonzero elements 82, ... , 8n of R, such that
(6.6) {82x2, ... , snxn } is a basis of N1
and
(6.7) si I s;+1 for 2 < i < n - 1.
Let S = S' U {x,}. Then the theorem is proved once we have shown tha
Si 132.
To verify that s1 $2, consider the element 82x2 E Ni C N and
1
let z = 81x1 + 32x2 E N. When we write z = c(z) z' where z' E M

is primitive and c(z) E R is the content of z, Remark 6.15 shows that
c(z) = (s1, 82). Thus, (Si) C (c(z)) and the maximality of (s1) in S shows
that (c(z)) = (s1), i.e., s1 I s2i and the proof of the theorem is complete. 0
(6.24) Example. Let N C Z2 be the submodule generated by y1 = (2, 4),

y2 = (2, -2), and y3 = (2, 10). Then c(y1) = c(y2) = c(y3) = 2. Further-
more, 2 divides every component of any linear combination of y1, y2, and
y3, so s1 = 2 in the notation of Theorem 6.23. Let v1 = (1, 2). Then
y1 = 2v1. Extend v1 to a basis of Z2 by taking v2 = (0,1). Then
(6.8) Ni = N n ((0, 1)) = ((0, 6)).
To see this note that every z E Ni can be written as
z = a1y1 + a2y2 + a3y3
where a1, a3, a3 E Z satisfy the equation
2a1 + 2a2 + 2a3 = 0.
Thus, 4a1 = -4a2 - 4a3, and considering the second coordinate of z, we

see that z = (z1, z2) where
z2 = 4a1 - 2a2 + 10a3 = -6a2 4. 6a3 = 6(a3 - a2).
Therefore, {v1i V2} is a basis of Z2, while {2v1, 6v2} is a basis of N. To

check, note that y1 = 2v1, y2 = 2v1 - 6v2, and y3 = 2v1 + 6v2.
(6.25) Remark. In Section 3.7, we will prove that the elements {s1, ... , sn}
are determined just by the rank n submodule N and not by the particular
choice of a basis S of M. These elements are called the invariant factors of
the submodule N in the free module M.
3.7 Finitely Generated Modules over PIDs

The invariant factor theorem for submodules (Theorem 6.23) gives a com-
plete description of a submodule N of a finitely generated free R-module
M over a PID R. Specifically, it states that a basis of M can be chosen so
that the first n = rank(N) elements of the basis, multiplied by elements
of R, provide a basis of N. Note that this result is a substantial general-
ization of the result from vector space theory, which states that any basis
of a subspace of a vector space can be extended to a basis of the ambient
space. We will now complete the analysis of finitely generated R-modules
(R a PID) by considering modules that need not be free. If the module M
is not free, then, of course, it is not possible to find a basis, but we will
still be able to express M as a finite direct sum of cyclic submodules; the
cyclic submodules may, however, have nontrivial annihilator. The following
result constitutes the fundamental structure theorem for finitely generated
modules over principal ideal domains.
(7.1) Theorem. Let M # 0 be a finitely generated module over the PID R.

If u(M) = n, then M is isomorphic to a direct sum of cyclic submodules
M-Rw1ED .®Rwn
such that
(7.1) R # Ann(wl) J Ann(w2) D . . . D Ann(wn) = Ann(M).
Moreover, for 1 < i < n
X7.2) Ann(wi) = Ann (M/(Rwi}1 + + Rwn)) .
Proof. Since µ(M) = n, let {vi, ... vn} be a generating set of M and
,
define an R-module homomorphism 0: Rn -+ M by

n
O(a1,
... , an) _ E aivi.

i=1
Let K = Ker(0). Since K is a submodule of Rn, it follows from Theorem

6.2 that K is a free R-module of rank in < n. By Theorem 6.23, there is a
basis {yl, ... , yn} of Rn and nonzero elements sl, ... , s,n E R such that
3.7 Finitely Generated Modules over PIDs 157
(7.3) {81y,, ... , smym} is a basis for K

and
(7.4) si I si+l for 1 < i < m - 1.
Let w; = ¢(yi) E M for 1 < i < n. Then {w1, ... ,wn} generates M
since 0 is surjective and {yl, ... , yn } is a basis of Rn. We claim that
By the characterization of direct sum modules (Theorem 3.1), it is sufficient

to check that if
(7.5) aiwi +.... + anwn = 0
where ai E R, then aiwi = 0 for all i. Thus suppose that Equation (7.5) is
satisfied. Then
ai (yl) + ... + anO(yn)

= -O(alyl + ... + anyn)
so that
a1y1 + ... +anyn E Ker(O) = K = (sly,, ... , smym)
Therefore,
aly, b1sly1
for some bl, ... , b,n E R. But {yl, ... , yn} is a basis of Rn, so we conclude
that a; = bisi for 1 < i < m while ai =0 form + 1 < i < n. Thus,
aiwi = bisiO(yi) = biO(siyi) = 0
for 1 < i < m because siyi E K = Ker(O), while aiwi = 0 for m+1 m, since (yi) fl Ker(O) = (0), it follows that Ann(wi) = (0).
Since si I sn for all i and since Ann(wi) = (si), we conclude that s,,M = 0.
Hence, Ann(wn) = (sn) = Ann(M) and Equation (7.1) is satisfied. Since
Equation (7.2) follows from Equation (7.1). The proof is now completed
by observing that Ann(wi) 54 R for any i since, if Ann(wi) = R, then
Rwi = (0), and hence, M could be generated by fewer than n elements.

But n = µ(M), so this is impossible because µ(M) is the minimal number
of generators of M.
A natural question to ask is to what extent is the cyclic decomposition

provided by Theorem 7.1 unique. Certainly, the factors themselves are not
unique as one can see from the example
Z2
Z.(1,0)ED
More generally, if M is a free R-module of rank n, then any choice of basis
{v1, ... ,v,a} provides a cyclic decomposition
Rv
with Ann(vi) = 0 for all i. Therefore, there is no hope that the cyclic factors
themselves are uniquely determined. What does turn out to be unique,
however, is the chain of annihilator ideals
Ann(wl) 2 ...
where we require that Ann(w;) # R, which simply means that we do not
allow copies of (0) in our direct sums of cyclic submodules. We reduce the
uniqueness of the annihilator ideals to the case of finitely generated torsion
R-modules by means of the following result. If M is an R-module, recall
that the torsion submodule M, of M is defined by
M, = {x E M : Ann(x) 34 (0)}-
(7.2) Proposition. If M and N are finitely generated modules over a PID

R, then Ma, N if and only if M, N, and rank M/M, = rank N/N,.
Proof. Let 0 : M -. N be an isomorphism. Then if x E M there is
an a 94 0 E R with ax = 0. Then aO(x) = O(ax) = 0(0) = 0 so that
O(x) E N,. Therefore, 0(M,) C N,. Applying the same observation to 0'
chows that O(M,) = N,. Thus, OIM, : M, -' N, is an isomorphism; if n :
N -. N/N, is the natural projection, it follows that Ker(ao0) = M,. The
first isomorphism theorem then gives an isomorphism M/M, = N/N,.
Since M/M, and N/N, are free R-modules of finite rank, they are isomor-
phic if and only if they have the same rank.
The converse follows from Corollary 6.20.
Therefore, our analysis of finitely generated R-modules over a PID R is

reduced to studying finitely generated torsion modules M; the uniqueness
of the cyclic submodule decomposition of finitely generated torsion modules
is the following result.
(7.3) Theorem. Let M be a finitely generated torsion module over a PID R,

and suppose that there are cyclic submodule decompositions
(7.6) MRwi
and
(7.7)
where
(7.8) Ann(wl) D Ann(wk) (0) with Ann(wl) 0 R
an d
(7.9) Ann(zl) D Ann(zr.) # (0) with Ann(zi) # R.
Then k = r and Ann(wi) = Ann(zi) for 1 < i < k.
Proof. Note that Ann(M) = Ann(wk) = Ann(zr). Indeed,
Ann(M) = Ann(Rwi + + Rwk)
= Ann(wi) fl fl Ann(wk)
= Ann(wk)
since Ann(wi) D Ann(wk). The equality Ann(M) = Ann(zr) is the
same argument.
We will first show that k = r. Suppose without loss of generality that
k > r. Choose a prime p E R such that (p) Ann(wi), i.e., p divides the
generator of Ann(wi). Then (p) D Ann(wi) for all i. Since p E Ann(M/pM),
it follows that M/pM is an R/pR-module and Equations (7.6) and (7.7)
imply
(7.10) MIpM Rwi/(pRwi) ®... Rwk/(pRwk)
and
(7.11) M/pM = Rzi/(pRzi) ® ... ® Rz,./(pRz,.).
Suppose that pRwi = Rwi. Then we can write apwi = wi for some a E R.
Hence, ap - 1 E Ann(wi) C (p) by our choice of p, so 1 E (p), which
contradicts the fact that p is a prime. Therefore, pRwi # Rwi for all i
and Equation (7.10) expresses the R/pR-module M/pM as a direct sum of
cyclic R/pR-modules, none of which is (0). Since R/pR is a field (in a PID
prime ideals are maximal), all R/pR-modules are free, so we conclude that
M/pM is free of rank k. Moreover, Equation (7.11) expresses M/pM as a
direct sum of r cyclic submodules, so it follows that k = µ(M/pM) < r.
Thus, r = k, and in particular, Rzi/(pRzi) 0 0 since, otherwise, M/pM
could be generated by fewer than k elements. Thus, (p) 2 Ann(i) for all
i; if not, then (p) + Ann(zi) = R, so there are a E R and c E Ann(zi) such
that ap + c = 1. Then zi = apzi + czi = apzi E pRzi, so Rzi/(pRzi) = 0,
and we just observed that Rzi/(pRzi) 54 0.
We are now ready to complete the proof. We will work by induction

on t(Ann(M)) where, if I = (a) is an ideal of R, then t(I) is the number
of elements (counted with multiplicity) in a prime factorization of a. This
number is well defined by the fundamental theorem of arithmetic for PIDs.
Suppose that t(Ann(M)) = 1. Then Ann(M) = (p) where p E R is prime.
Since Ann(M) = Ann(wk) = Ann(zk) = (p) and since (p) is a maximal
ideal, Equations (7.8) and (7.9) imply that Ann(wi) = (p) = Ann(zi) for
all i, and the theorem is proved in the case t(Ann(M)) = 1.
Now suppose the theorem is true for all finitely generated torsion R-
modules N with t(Ann(N)) < t(Ann(M)), and consider the isomorphisms
(7.12) pM'5 pRwl ® ... ® pRwk pRw,+i ® ... ® pRwk
and
(7.13) pM L' pRzl ® ... ® pRzk =' pRzt+i ® ... ® pRzk
where Ann(wj) = ... = Ann(w,) = Ann(zi) = ... = Ann(zi) = (p) and
Ann(w,+i) g6 (p), Ann(zt+i) 0 (p) (s and t may be 0). Then Ann(pM) =
(a/p) where Ann(M) = (a), so t(Ann(pM)) = t(Ann(M))-1. By induction
we conclude that k - s = k - t, i.e., s = t, and Ann(pwi) = Ann(pzi) for
s < i < k. But Ann(pwi) = (ai/p) where Ann(wi) = (ai). Thus Ann(wi) _
Ann(zi) for all i and we are done. 0
Since Rwi R/ Ann(wi) and since R/I and R/J are isomorphic R-
modules if and only if I = J (Exercise 10), we may rephrase our results as
follows.
(7.4) Corollary. Finitely generated modules over a PID R are in one-to-one

correspondence with finite nonincreasing chains of ideals
R#I,212_J...2In.
Such a chain of ideals corresponds to the module
M=R/Il®...®R/In.
Note that µ(M) = n and if Ik+i In = (0) but Ik 96 (0), then
M^_'R/II®...®R/Ik®R"-k
We will use the convention that the empty sequence of ideals (n = 0) cor-
responds to M = (0).
Proof 0
(7.5) Definition. If M is a finitely generated torsion module over a PID R

and M Rwi ® ®Rwn with Ann(wi) Ann(wi+i) (1 < i < n - 1)
and Ann(wi) 0 R, then the chain of ideals Ii = Ann(wi) is called the chain
of invariant ideals of M.
3.7 Finitely Generated Modules over PII)s 161
Using this language, we can express our results as follows:
(7.6) Corollary. Two finitely generated torsion modules over a PID are iso-
morphic if and only if they have the same chain of invariant ideals.
Proof. 0
(7.7) Remark. In some cases the principal ideals Ann(w3) have a preferred
generator aj. In this case the generators {ajIj=1 are called the invariant
factors of M.
The common examples are R = Z, in which case we choose aj > 0 so

that aj = IZ/Ann(w,)I, and R = F[X], where we take monic polynomials
as the preferred generators of ideals.
(7.8) Definition. Let R be a PID, and let M be a finitely generated torsion

R-module with chain of invariant ideals
(31) ? (92) ? ... D (3n)-
We define me(M) = sn and co(M) = 81 sn.
Note that me(M) and co(M) are only defined up to multiplication by a

unit, but in some cases (R = Z or R = F[X]) we have a preferred choice of
generators of ideals. In these cases me(M) and co(M) are uniquely defined.
Concerning the invariants me(M) and co(M), there is the following trivial
but useful corollary of our structure theorems.
(7.9) Corollary. Let M be a finitely generated torsion module over a PID

R.
(1) If a E R with aM = 0, then me(M) I a.
(2) me(M) divides co(M).
(3) If p E R is a prime dividing co(M), then p divides me(M).
Proof. (1) Since Ann(M) = (sn) = (me(M)) by Theorem 7.1 and the
defintion of me(M), it follows that if aM = 0, i.e., a E Ann(M), then
me(M) I a.
(2) Clearly s,, divides s1 ... sn.
(3) Suppose that p I 31 sn = co(M). Then p divides some si, but
(si) 2 (sn), so si I sn. Hence, p I sn = me(M). 0
(7.10) Remark. There are, unfortunately, no standard names for these in-
variants. The notation we have chosen reflects the common terminology in
the two cases R = Z and R = F[X]. In the case R = Z, me(M) is the
exponent and co(M) is the order of the finitely generated torsion Z-module
(= finite abelian group) M. In the case R = FIX] of applications to lin-

ear algebra to be considered in Chapter 4, me(VT) will be the minimal
polynomial and co(VT) will be the characteristic polynomial of the linear
transformation T E HomF (V) where V is a finite-dimensional vector space
over the field F and VT is the FIX]-module determined by T (see Example
1.5 (12)).
There is another decomposition of a torsion R-module M into a direct

sum of cyclic submodules which takes advantage of the prime factorization
of any generator of Ann(M). To describe this decomposition we need the
following definition.
(7.11) Definition. Let M be a module over the PID R and let p E R be a

prime. Define the p-component Mp of M by
M p = {x E M : Ann(x) = (p") for some n E Z+}.
If M = Mp, then M is said to be p-primary, and M is primary if it is
p-primary for some prime p E R.
It is a simple exercise to check that submodules, quotient modules, and

direct sums of p-primary modules are p-primary (Exercise 54).
(7.12) Theorem. If M is a finitely generated torsion module over a PID R,

then M is a direct sum of primary submodules.
Proof. Since M is a direct sum of cyclic submodules by Theorem 7.1, it
is sufficient to assume that M is cyclic. Thus suppose that M = (x) and
suppose that
Ann(x) = (a) = (pr' ... pr,,)
where pl, ... , pn are the distinct prime divisors of a. Let qi = a/p;'. Then
1 = (ql, ... , qn) = gcd{ql, ... , q,,}, so there are bl, ... , bn E R such that
(7.14) 1 = bigi + ... + bngn
Let xi = bigix. Then Equation (7.14) implies that
x=xl +...+ xn
so that
Al = (x) (XI) + ... + (xn)
Suppose that y E (x1) fl ((x2) + + (x,)). Then
C2x2+...+Cnxn
y=Clx1
and hence, p 'y = clb1Pi'gix = c1blax = 0 and
qly = C2g2P23x2 + +CngnPn"xn = 0,
where q = ql Therefore, {pl' , ql } C Ann(y), but (pi' , ql) = 1 so that

Ann(y) = R. Therefore, y = 0. A similar calculation shows that
(xi) n ((xl) + ... + (xi) + ... + (xn)) = (0),
so by Theorem C
Combining Theorems 7.1 and 7.12, we obtain the following result:
(7.13) Theorem. Any finitely generated torsion module M over a PID R is

a direct sum of primary cyclic submodules.
Proof. Suppose M Rwl® . ®Rwn as in Theorem 7.1. Then if Ann(wi) _
(si), we have si I si+1 for 1 < i < n - 1 with sl 54 1 and sn # 0 (since M
is torsion). Let p', ... pk be the set of distinct nonassociate primes that
occur as a prime divisor of some invariant factor of M. Then
Si = ulpil, ... peklk
pknk
sn = unplnl ...
where the divisibility conditions imply that
0 < elf < e2j < ... < eni for 1<j<k.
Then the proof of Theorem 7.12 shows that M is the direct sum of cyclic
submodules with annihilators eij > 0}, and the theorem is proved.
0
(7.14) Definition. The prime powers eij > 0, 1 < j < k} are called
the elementary divisors of M.
(7.15) Theorem. If M and N are finitely generated torsion modules over a

PID R, then M N if and only if M and N have the same elementary
divisors.
Proof. Since M is uniquely determined up to isomorphism from the invari-
ant factors, it is sufficient to show that the invariant factors of M can be
recovered from a knowledge of the elementary divisors. Thus suppose that
... (sn)
(sl) (32)
is the chain of invariant ideals of the finitely generated torsion module M.

This means that si ( si+1 for 1 < i < n. Let pl, ... ,pk be the set of distinct
nonassociate primes that occur as a prime divisor of some invariant factor
of M. Then
$1 = U1P11" ... Pkik

(7.15)
P`"k
sn = unP1"' ... k
where the divisibility conditions imply that
(7.16) 0 < e1j < e2j < < enj for 1 < j < k.
Thus, the elementary divisors of M are

(7.17)
We show that the set of invariant factors (Equation (7.15)) can be recon-
structed from the set of prime powers in Equation (7.17). Indeed, if
ej= maxe;j,
1<.<n
1<j<k,
then the inequalities (7.16) imply that sn is an associate of pi' pkk. Delete
(Pi' , ... , Pkk }
from the set of prime powers in set (7.17), and repeat the process with
the set of remaining elementary divisors to obtain sn_1. Continue until all
prime powers have been used. At this point, all invariant factors have been
recovered. Notice that the number n of invariant factors is easily recovered
from the set of elementary divisors of M. Since s1 divides every si, it follows
that every prime dividing 81 must also be a prime divisor of every s,.
Therefore, in the set of elementary divisors, n is the maximum number of
occurrences of p"i for a single prime p. 0
(7.16) Example. Suppose that M is the Z-module

M = Z22 X Z22 X Z3 X Z32 X Z5 X Z7 X Z7a .
Then the elementary divisors of M are 22, 22, 3,3 2 , 5, 7,7 2. Using the
algorithm from Theorem 7.15, we can recover the invariant factor descrip-
tion of M as follows. The largest invariant factor is the product of the
highest power of each prime occurring in the set of elementary divisors,
i.e., the least common multiple of the set of elementary divisors. That is,
32 = 72 5 32 22 = 8820. Note that the number of invariant factors of
M is 2 since powers of the primes 2, 3, and 7 occur twice in the set of ele-
mentary divisors, while no prime has three powers among this set. Deleting
72, 5,3 2 , 22 from the set of elementary divisors, we obtain s1 = 7.3.22 = 84.
This uses all the elementary divisors, so we obtain
Al Z&4 X Zs82o
We now present some useful observations concerning the invariants

me(M) and co(M) where M is a torsion R-module (R a PID). See Definition
7.8 for the definition of these invariants. The verification of the results that
we wish to prove require some preliminary results on torsion R-modules,
which are of interest in their own right. We start with the following lemmas.
(7.17) Lemma. Let M be a module over a PID R and suppose that x E MT.
If Ann(x) = (r) and a E R with (a, r) = d (recall that (a, r) = gcd{a, r}),
then Ann(ax) = (r/d).
Proof. Since (r/d)(ax) = (a/d)(rx) = 0, it follows that (r/d) C Ann(ax).
If b(ax) = 0, then r I (ba), so ba = rc for some c E R. But (a, r) = d, so
there are s, t E R with rs + at = d. Then rct = bat = b(d - rs) and we see
that bd = r(ct + bs). Therefore, b E (r/d) and hence Ann(ax) = (r/d). 0
(7.18) Lemma Let M be a module over a PID R, and let x1, ... , xn E Mr
with Ann(xi) = (ri) for 1 < i < n. If {rl, ... , rn} is a pairwise relatively
prime subset of R and x = x1 + + xn, then Ann(x) _ (a) _ (f 1 ri).
Conversely, if y E M, is an element such that Ann(y) _ (b) si)
where {s1, . . . , s, } is a pairwise relatively prime subset of R, then we may
write y = yl + + yn where Ann(yi) _ (si) for all i.
Let x = x1 + - + xn. Then a = rj 1 ri E Ann(x) so that (a) C
P r o o f.
Ann(x). It remains to check that Ann(x) C (a). Thus, suppose that bx = 0.
By the Chinese remainder theorem (Theorem 2.2.24), there are c1, ... , cn E
R such that
1 (mod (ri)),
ci = 0 (mod (r3)), if j 0 i.
Then, since (rj) = Ann(xj), we conclude that cixj = 0 if i 96 j, so for each
iwith 1<i<n
Therefore, bci E Ann(xi) = (ri), and since ci - 1 (mod (ri)), it follows that
ri I b for 1 < i < n. But {ri, ... , rn } is pairwise relatively prime and thus
Conversely, suppose that y E M satisfies Ann(y) = (b) = (i=1 f)

a is the least common multiple of the set {r1, ... ,rn}. We conclude that
a I b, and hence, Ann(x) = (a).
where the set {s1, ... , sn } is pairwise relatively prime. As in the above
paragraph, apply the Chinese remainder theorem to get c1, ... , en E R
such that
(I (mod (si)),
c' = 0 (mod (sj)), if j , i.
Since b is the least common multiple of {s1i ... , sn}, it follows that
1 - c1 (mod (b)),
and hence, if we set yi = cjy, we have

y1+...+yn=(c1+...+cn)y=y
Since (b, c1) sj), Lemma 7.17 shows that Ann(yi) = Ann(ciy) _
(se). 0
(7.19) Proposition. Let R be a PID and suppose that M is a torsion R-
module such that
M Rw1 ®... Rwn
with Ann(wi) = (ti). Then the prime power factors of the t; (1 < i < n)
are the elementary divisors of M.
Proof. Let pi, ... pk be the set of distinct nonassociate primes that occur
as a prime divisor of some ti. Then we may write
pklk
t1 = ulpi" ...
(7.18)
to = unpl. ...
where u1i ... , un are units in R and some of the exponents eij may be 0.
The proof of Theorem 7.12 shows that
Rwi Rzi1 ® ... Rzik
where Ann(zij) For notational convenience we are allowing zij = 0
for those (i, j) with eij = 0. Therefore,
(7.19) Al f___ (3) Rzjj

is
where Ann(zij) = Let S = {pe'' } where we allow multiple occur-
rences of a prime power pe, and let
S = {Zij}.
Let m be the maximum number of occurrences of positive powers of a single
prime in S. If
(7.20) fnj = max eij for 1 < j < k,
1 0 for 1 < j < k.
Delete {pf-1, ... , pf-* } from the set S and repeat the above process
with the remaining prime powers until no further positive prime powers are
available. Since a prime power for a particular prime p is used only once at
each step, this will produce elements al, ... , sm E R. From the inductive
description of the construction of si, it is clear that every prime dividing s;
also divides s;+1 to at least as high a power (because of Equation (7.21)).
Thus,
s; s;+1 for 1 0}={pWs:eo,6 > 0}
where repetitions of prime powers are allowed and where
(7.24) 0:5 for 1<j<k
by Equation (7.20).
For each (1 < i < m), choose w;j E S with Ann(w;j) =
and let xi = wit + + wik. Lemma 7.18 shows that Ann(x;) = (si) for
1 < i < m, and thus,
k k
Rx; Rl (si) °-` ®Rl (Pf ") ®Rw:j
j=1 j=1
Equation (7.19) then shows that
M ®RzaQ
0.0
m k
?® Rw;j
i=1 j=1
! Rxl®...0Rzm
where Ann(x;) = (s;). Since si s;+1 for 1 < i < m, it follows that
I
{Si, ... , am} are the invariant factors of M, and since the set of prime
power factors of {Si, ... , sm} (counting multiplicities) is the same as the
set of prime power factors of {t1i ... , t,,} (see Equation (7.23)), the proof
is complete.
(7.20) Corollary. Let R be a PID, let M1, ... , Mk be finitely generated

torsion R-modules, and let M = ®k ,M;- If {d;1, ... , die;} } is the set of
elementary divisors of Mi, then
S={d,,:1<i<k;I<j<ti}
is the set of elementary divisors of M.
Proof. By Theorem 7.1,
Mi ?' Rwi1 ®... ® Rwi,.,

where Ann(wi1) = (sib) and si, si,j+l for 1 < j < ri. The elementary
I
divisors of Mi are the prime power factors of {sit, ... , si,., }. Then
k
M=®MiL, ®Rwij
i=1 i,j
where Ann(wij) = (sjj). The result now follows from Proposition 7.19. O
(7.21) Proposition. Let R be a PID, let M1, ... , Mk be finitely generated

torsion R-modules, and let M = ®;` 1Mi. Then
(7.25) me(M) = lcm{me(M1), ... ,me(Mk))
k
(7.26) co(M) _ [Ico(Ms).
i=1
Proof. Since Ann(M) = Ik 1 Ann(MM), Equation (7.25) follows since

I
(me(Mi)) = Ann(M5). Since co(M) is the product of all invariant factors of

M, which is also the product of all the elementary divisors of M, Equation
(7.26) follows from Corollary 7.20. 0
The special case R = Z is important enough to emphasize what the

results mean in this case. Suppose that M is an abelian group, i.e., a Z-
module. Then an element x E M is torsion if and only if nx = 0 for
some n > 0. That is to say, x E MT if and only if o(x) < oo. Moreover,
Ann(x) = (n) means that o(x) = n. Thus the torsion submodule of M
consists of the set of elements of finite order. Furthermore, M is finitely
generated and torsion if and only if M is a finite abelian group. Indeed, if
M = (x1, ... , xk) then any x E M can be written x = nix1 + + nkxk
where 0 < ni < o(xi) < oo f o r 1 < i < k. Therefore, IMI < n 1 o(xi).
Hence, the fundamental structure theorem for finitely generated abelian
groups takes the following form.
(7.22) Theorem. Any finitely generated abelian group M is isomorphic to

Zr ®M1 where IMi I < oo. The integer r is an invariant of M. Any finite
abelian group is a direct sum of cyclic groups of prime power order and
these prime power orders, counted with multiplicity, completely characterize
the finite abelian group up to isomorpism. Also any finite abelian group is
uniquely isomorphic to a group
Z,, X . . X Z,k
where 8i 1 si+i for all i.

Proof.
Given a natural number n it is possible to give a complete list of all

abelian groups of order n, up to isomorphism, by writing n = pi' Pkk
where pl, ... ,pk are the distinct prime divisors of n. Let M be an abelian
group of order n. Then we may write M as a direct sum of its primary
components
M^='MM ED ...ED Mpk
where IMp, I Then each primary component Mp, can be written as a
direct sum
Mp,?'Zp'-,,® ®Zp:u
where
1<ej,:5 e;t<ri
and
Furthermore, the main structure theorems state that M is determined up

to isomorphism by the primes pl, ... , pk and the partitions ei1, ... , eit of
the exponents r;. This is simply the statement that M is determined up to
isomorphism by its elementary divisors. Therefore, to identify all abelian
groups of order n, it is sufficient to identify all partitions of ri, i.e., all ways
to write r; = e;1 + + e;t as a sum of natural numbers.
(7.23) Example. We will carry out the above procedure for n = 600 =
23 3.52. There are three primes, namely, 2, 3, and 5. The exponent of 2
is3andwecanwrite 3=1+1+1,3=1+2,and3=3.Thus there are
three partitions of 3. The exponent of 3 is 1, so there is only one partition,
while the exponent of 5 is 2, which has two partitions, namely, 2 = 1 + 1
and 2 = 2. Thus there are 3.1.2 = 6 distinct, abelian groups of order 600.
They are
Z2 X Z2 X Z2 X Z3 X Z5 X Z5 L Z2 X Z10 X Z30
Z2 X Z2 X Z2 X Z3 X Z25 = Z2 X Z2 X Z150
Z2XZ4XZ3XZ5XZ5~Z10XZ60
Z2 X Z4 X Z3 X Z25 = Z2 X Z300
z8xZ3xZ5xZ5Z5XZ120
z8xZ3XZ25Z600
where the groups on the right are expressed in invariant factor form and
those on the left are decomposed following the elementary divisors.
We will conclude this section with the following result concerning the
structure of finite subgroups of the multiplicative group of a field. This
is an important result, which combines the structure theorem for finite
abelian groups with a bound on the number of roots of a polynomial with
coefficients in a field.
(7.24) Theorem. Let F be a field and let G C F' = F \ {0} be a finite

subgroup of the multiplicative group F. Then G is a cyclic group.
Proof. According to Theorem 7.1, G is isomorphic to a direct sum
G(z1)ED ...ED (zn)
where, if we let ki = o(zi) = order of z,, then ki I ki+1 for 1 < i 5 n - 1
and
Ann(G) = Ann(zn) _ (kn)Z.
In the language of Definition 7.8, me(G) = kn. This means that zk.. =I
for all z E G. Now consider the polynomial
(7.27) P(X) = Xk^ - 1.
Since F is a field, the polynomial P(X) has at most kn roots, because degree
P(X) = kn (Corollary 2.4.7). But, as we have observed, every element of
G is a root of P(X), and
JGJ = k1k2 ... kn.
Thus, we must have n = 1 and G = (z1) is cyclic.
(7.25) Corollary. Suppose that F is a finite field with q elements. Then F'
is a cyclic group with q - 1 elements, and every element of F is a root of
the polynomial X9 - X.
Proof. Exercise.
(7.28) Corollary. Let

Gn = {e2xi(k/n) : 0 < k < n- 1} C C.
Then Gn is the only subgroup of C' of order n.
3.8 Complemented Submodules 171
Proof. Let H be a finite subgroup of C' with IHJ = n. Then every element
z of H has the property that z" = 1. In other words, z is a root of the
equation X" = 1. Since this equation has at most n roots in C and since
every element of G,, is a root of this equation, we have z E G". Thus, we
conclude that H C G" and hence H = G" because n = IHI = IG"I. 0
3.8 Complemented Submodules

We will now consider the problem of extending a linearly independent sub-
set of a free R-module to a basis. The example {2} C Z shows that some
restrictions on the subset are needed, while Theorem 6.16 shows that any
primitive element of a finitely generated free R-module (R a PID) can be
extended to a basis.
(8.1) Definition. Let M be an R-module and S C M a submodule. Then

S is said to be complemented if there exists a submodule T C M with
M be a finitely generated free R-module with basis {v1, ... , v,,)

and let S = (v1, ... , v,). Then S is complemented by T = (v,+1, ... , v").
This example shows that if W = {w1, ... wk} is a linearly independent
subset of M, then a necessary condition for W to extend to a basis of M is
that the submodule (W) be complemented. If R is a PID, then the converse
is also true. Indeed, let T be a complement of (W) in M. Since R is a PID,
T is free, so let {x1, ... , x,.} be a basis of T. Then it is easy to check that
{w1,...,wk,x1,...,x,.}is a basis of M.
(8.2) Proposition. Let R be a PID, let M be a free R-module, and let S be

a submodule. Consider the following conditions on S.
(1) S is complemented.
(2) MIS is free.
(3) IfxESandx=ayforsomeyEM,aO0ER,thenyES.
Then (1) s (2) and (2) (3), while if M is finitely generated, then
(3) = (1).
Proof. (1) (2). If S is complemented, then there exists T C M such that

S E T Q5 M. Thus, M/S se T. But T is a submodule of a free module over
a PID R, so T is free (Theorem 6.2).
(2) =: (3). Suppose MIS is free. If x E S satisfies x = ay for some
y E M, a 0 0 E R, then a(y + S) = S in MIS. Since free modules are
torsion-free, it follows that y + S = S, i.e., y E S.
(3) = (1). Let M be a finite rank free R-module and let S C M be a

submodule satisfying condition (3). Then there is a short exact sequence
(8.1) O-S- M M/S- O
where n is the projection map. Condition (3) is equivalent to the statement
that MIS is torsion-free, so M/S is free by Theorem 6.6. But free modules
are projective, so sequence (8.1) has a splitting or : MIS -. M and Theorem
3.9 shows that M - S ® Im(a), i.e., S is complemented.
(8.3) Remarks.
(1) A submodule S of M that satisfies condition (3) of Proposition 8.2
is called a pure submodule of M. Thus, a submodule of a finitely
generated module over a PID is pure if and only if it is complemented.
(2) If R is a field, then every subspace S C M satisfies condition (3) so that
every subspace of a finite-dimensional vector space is complemented.
Actually, this is true without the finite dimensionality assumption, but
our argument has only been presented in the more restricted case. The
fact that arbitrary subspaces of vector spaces are complemented follows
from Corollary 4.21.
(3) The implication (3) (1) is false without the hypothesis that M be
finitely generated. As an example, consider a free presentation of the
Z-module Q:
O-*S-M-+Q-0.
Since MIS - Q and Q is torsion-free, it follows that S satisfies con-
dition (3) of Proposition 8.1. However, if S is complemented, then a
complement T - Q; so Q is a submodule of a free Z-module M, and
hence Q would be free, but Q is not a free Z-module.
(8.4) Corollary. If S is a complemented submodule of a finitely generated

R-module (R a PID), then any basis for S extends to a basis for M.
Proof. This was observed prior to Proposition 8.2.
(8.5) Corollary. If S is a complemented submodule of M, then rank S =

rank M if and only if S = M.
Proof. A basis {vl, ... , v,n} of S extends to a basis {vl, ... , vn} of M. But
n=m,soS=(vl,...,vn)=M.
If AI = Z and S = (2), then rank S = rank M but M 0 S. Of course,
S is not complemented.
(8.6) Corollary. If S is a complemented submodule of M, then

rank M = rank S + rank(M/S).
3.8 Complemented Submodules 173
Proof. Let S = (v1, ... , v,,,) where m = rank S. Extend this to a basis
{v1, ... , of M. Then T = (v,,,+i, ... , v,,) is a complement of S in M
and T '_5 M/S. Thus,
rank M=n=m+(n-m)=rank S+rank(M/S).
(8.7) Proposition. Let R be a PID and let f : M -. N be an R-module
homomorphism of finite-rank free R-modules. Then
(1) Ker(f) is a pure submodule, but
(2) Im(f) need not be pure.
Proof. (1) Suppose X E Ker(f), a # 0 E R, and y E M with x = ay. Then

0 = f (x) = f (ay) = a f (y). But N is free and, hence, torsion-free so that
f (y) = 0. Hence, condition (3) of Proposition 8.2 is satisfied, so Ker(f) is
complemented.
(2) If f : Z -. Z is defined by f (x) = 2x, then Im(f) = 2Z is not a
complemented submodule of Z.
(8.8) Proposition. Let R be a PID and let f : M - N be an R-module

homomorphism of finite-rank free R-modules. Then
rank M = rank(Ker(f )) + rank(Im(f )).
Proof. By the first isomorphism theorem, Im(f) M/ Ker(f ). But Ker(f)

is a complemented submodule of M, so the result follows from Corollary
8.6.
(8.9) Corollary. Let R be a PID and let M and N be finite-rank free

R-modules with rank(M) = rank(N). Let f E HomR(M, N).
(1) If f is a surjection, then f is an isomorphism.
(2) If f is an injection and Im(f) is complemented, then f is an isomor-
phism.
Proof. (1) By Proposition 8.8, rank(Ker(f)) = 0, i.e., Ker(f) = (0), so f is

an injection.
(2) By Proposition 8.8, rank N = rank M = rank(Im(f )). Since Im(f)
is complemented by hypothesis, f is a surjection by Corollary 8.5.
(8.10) Proposition. Let R be a field and let M and N be R-modules with

rank(M) = rank(N) finite. Let f E HomR(M, N). Then the following are
equivalent.
(1) f is an isomorphism.
(8.11) Proposition. Let M be a finite-rank free R-module (R a PID). If

S and T are pure submodules, then
(2) f is an injection.
(3) f is a surjection.
Proof. Since R is a field, Im(f) is complemented (by Remark 8.3 (2)), so

this is an immediate consequence of Corollary 8.9. 0
(8.11) Proposition. Let M be a finite-rank free R-module (R a PID). If S
and T are pure submodules, then
rank(S + T) + rank(S n T) = rank S + rank T.
Proof Note that if S and T are pure submodules of M, then S n T is also

pure. Indeed, if ay E S n T with a 0 0 E R, then y E S and y E T since
these submodules are pure. Thus, y E S n T, so S n T is complemented by
Proposition 8.2. Then
(S + T)/T S/(S n T).
By Corollary 8.6, we conclude
rank(S + T) - rank(T) = rank(S) - rank(S n T).
0
(8.12) Remark. It need not be true that S + T is pure, even if S and T are
both pure. For example, let S = ((1, 1)) C Z2 and let T = ((1, -1)) C Z2.
Then S and T are both pure, but S+T 76 Z2, so it cannot be pure. In fact,
2.(1,0)=(2,0)=(1, 1)+(1, -1) ES+T,
3.9 Exercises
1. If M is an abelian group, then Endz(M), the set of abelian group endomor-
phisms of M, is a ring under addition and composition of functions.
(a) If M is a left R-module, show that the function 0 : R -+ Endz(M)
defined by 0(r)(m) = rm is a ring homomorphism. Conversely, show
that any ring homomorphism 0 : R -. Endz(M) determines a left R-
module structure on M.
(b) Show that giving a right R-module structure on M is the same as giving
a ring homomorphism 0: R°' -+ Endz(M).
2. Show that an abelian group G admits the structure of a if and
only if nG = (0).
3. Show that the subring Z(4 ] of Q is not finitely generated as a Z-module if
9 ¢ Z.
4. Let M be an S module and suppose that R C S is a subring. Then 1Li is also
an R-module by Example 1.5 10). Suppose Chat N C M is an R-submodule.
LetSN=(sn:sES, nEN .
(a) If S = and R = Z, show that SN is the S-submodule of M generated
by N.
(b) Show that the conclusion of part (a) need not hold if S = R and R = Q.
3.9 Exercises 175
5. Let M be an R-module and let A, B, and C be submodules. If C C A, prove

that
An(B+C)=(AnB)+C.
This equality is known as the modular law. Show, by example, that this
formula need not hold if C is not contained in A.
6. Let R be a commutative ring and let S C R\ {0} be a multiplicatively closed
subset of R containing no zero divisors. Let M be an R-module. Mimicking
the construction of RS (Theorem 2.3.5), we define MS as follows. Define a
relation - on M x S by setting (x, s) - (y, t) if and only if utx = usy
for some u E S. Verify that this is an equivalence relation (see the proof
of Theorem 2.3.5). We will denote the equivalence class of (x, s) by the
suggestive symbol x/s.
a Prove that MS is an Rs-module via the operation (a/s)(x/t) = (ax)/(st).
(b) If f : M -* N is an R-module homomorhism, show that Is : Ms NS
defined by fs(x/s) = f(x)/s is an RS-module homomorphism.
(c) If x E M, show that x/I = 0 in MS if and only if Ann(x) n S 76 0.
7. Let R C FIX] be the subring

R = {f(X) E FIX] : f(X) =ao+a2X2+...+anX"}.
Thus, f (X) E R if and only if the coefficient of X is 0. Show that FIX] is a
finitely generated R-module that is torsion-free but not free.
8. Show that Q is a torsion-free Z-module that is not free.
9. (a) Let R be an integral domain, let M be a torsion R-module, and let N
be a torsion-free R-module. Show that HomR M, N) = (0).
(b) According to part (a), Homz(Z,n, Z) = (0). If n = km, then Z,n is a
Zn-module. Show that
Zn) °-` Zm-
10. Let R be a commutative ring with 1 and let I and J be ideals of R. Prove
that R/I °_° R/J as R-modules if and only if I = J. Suppose ask
X
that R/I and /J be isomorphic rings. Is the same conclusion valid?
id? (Hint:
Consider F[X]/(X - a) where a E F and show that F[X]/(X - a) °_° F as
rings.)
11. Prove Theorem 2.7.
12. Prove Lemma 2.9.
13. Let M be an R-module and let f E EndR(M) be an idempotent endomor-
phism of M, i.e., f o f = f. (That is, f is an idempotent element of the ring
EndR(M).) Show that
M °- (Ker(f)) ® (Im(f)).
14. Prove the remaining cases in Theorem 3.10.
15. Let R be a PID and let a and b E R be nonzero elements. Then show
that HomR(R/Ra, R/Rb) ?f R/Rd where d = (a, b) is the greatest common
divisor of a and b.
16. Compute Homz(Q, Z).
17. Give examples of short exact sequences of R-modules
and
O-.N1 -.N
such that
a) Mi'`N1, M2tN, M2 N2;
b) Mi° NI,M N,M22.N2;
c) M1 N1, M °__ N, M2 °5 N2
18. Show that there is a split exact sequence
0 - . mZmn - Zmn -+ nZmn - 0
of Zmn-modules if and only if (m, n) = 1.
19. Let N1 and N2 be submodules of an R-module M. Show that there is an
exact sequence
O-.N1nN2!N1®N2-N1+N2-+0
where p(x) = (x, x) and m(x, y) = x - y.
20. Let R be an integral domain and let a and b be nonzero elements of R. Let
M = R/R(ab) and let N = Ra/R(ab). Then M is an R-module and N is a
subm ule. Show that N is a complemented submodule in M if and only if
there are u, v E R such that ua + vb = I.
21. Let R be a ring, M a finitely generated R-module, and 4 : M -+ R" a
surjective R-module homomorphism. Show that Ker ¢) is finitely generated.
(Note that this is valid even when M has submodules that are not finitely
generated.) (Hint: Consider the short exact sequence:
0 -+ K -. M m-+ R" -+ 0. )
22. Suppose that
0 , M1 m M - M2 -+ 0
0 -. Ni1' 19
N
1h
N2 -0
is a commutative diagram of R-modules and R-module homomorphisms.
Assume that the rows are exact and that f and h are isomorphisms. Then
prove that g is an isomorphism.
23. Let R be a commutative ring and S a multiplicatively closed subset of R
containing no zero divisors. If M is an R-module, then Ms was defined in
Exercise 6. Prove that the operation of forming quotients with elements of
S is exact. Precisely:
(a) Suppose that M' f M 9+ M" is a sequence of R-modules and homo-
morphisms which is exact at M. Show that the sequence
Ms Is Ms gs MS
is an exact sequence of Rs-modules and homomorphisms.
(b) As a consequence of part (a), show that if M' is a submodule of M, then
Ms can be identified with an Rs-submodule of Ms.
(c) If N and P are R-submodules of M, show (under the identification
of part (b)) that (N + P)s = Ns + Ps and (N n P)s = Ns n Ps.
(That is, formation of fractions commutes with finite sums and finite
intersections.)
(d) If N is a submodule of M show that
3.9 Exercises 177
(M/N)s'-` (Ms)/(Ns)-(That
is, formation of fractions commutes with quotients.)

24. Let F be a field and let {f;(X)},:o be any subset of F[X] such that
degf;(X) = i for each i. Show that {f;(X)} o is a basis of FIX) as an
F-module.
25. Let R be a commutative ring and consider M = R[X] as an R-module. Then
N = R[X21 is an R-submodule. Show that M/N is isomorphic to R[X] as
an R-module.
26. Let G be a group and H a subgroup. If F is a field, then we may form the
group ring F(G) (Example 2.1.9 (15)). Since F(G) is a ring and F(H) is
a subring, we may consider F(G) as either a left F(H)-module or a right
F(H)-module. As either a left or right F( H)-module, show that F(G) is free
of rank [G : H]. (Use a complete set {g; of coset representatives of H as a
basis.)
27. Let R and S be integral domains and let 01, ... , ¢, be n distinct ring
homomorphisms from R to S. Show that are S-linearly indepen-
dent in the S-module F(R, S) of all functions from R to S. (Hint: Argue by
induction on n, using the property O;(ax) = O;(a)O;(x), to reduce from a
dependence relation with n entries to one with n - 1 entries.)
28. Let G be a group, let F be a field, and let O; : G -. F' for 1 < i < n
be n distinct group homomorphisms from G into the multiplicative group
F' of F. Show that 01, ..., 0n are linearly independent over F (viewed as
elements of the F-module of all functions from G to F). (Hint: Argue by
induction on n, as in Exercise 27.)
29. Let R = Z3o and let A E M2,3(R) be the matrix
A= 1 1 -11 .
0 2 3
Show that the two rows of A are linearly independent over R, but that any
two of the three columns are linearly dependent over R.
30. Let V be a finite-dimensional complex vector space. Then V is also a vector
space over R. Show that dime, V = 2dimc V. (Hint: If
B={vi,...,vn}
is a basis of V over C, show that
i3' = {vl, ... vn, ivl, ... ivn)
is a basis of V over R.)
31. Extend Exercise 30 as follows. Let L be a field and let K be a subfield of L.
If V is a vector space over L, then it is also a vector space over K. Prove
that
dimK V = [L: K] dim,. V
where [L : K] = dimK L is the dimension of L as a vector space over K.
(Note that we are not assuming that dimK L < oo.)
32. Let K C L be fields and let V be a vector space over L. Suppose that
B = {ua },Er is a basis of V as an L-module, and let W be the K-submodule
of V generated by B. Let U C W be any K-submodule, and let U,. be the
L-submodule of V generated by U. Prove that
U,.nW=U.
That is, taking L-linear combinations of elements of U does not produce any
new elements of W.
That is, taking L-linear combinations of elements of U does not produce any
new elements of W.
33. Let K C L be fields and let A E A, (K), b E M.,1(K). Show that the matrix
equation AX = b has a solution X E M,,,1(K) if and only if it has a solution
X E M,,,I(L).
34. Prove that the Lagrange interpolation polynomials (Proposition 2.4.10) and
the Newton interpolation polynomials (Remark 2.4.11) each form a basis of
the vector space of polynomials of degree < n with coefficients from
F.
35. Let F denote the set of all functions from Z+ to Z+, and let M be the
free Q-module with basis F. Define a multiplication on M by the formula
(f g)(n) = f (n) + g(n) for all f, g E F and extend this multiplication by
linearity to all of M. Let fm be the function f,,, n) = 6.,,,,, for all m, n > 0.
Show that each fm is irreducible in fact, prime) as an element of the ring
onsider the function f (n) = 1 for all it > 0. Show that f does not
Al. Now consider'
have a factorization into irreducible elements in Al. (Hint: It may help to
think of f as the "infinite monomial"
Xo ro)Xi (1) ... Xm(-) ....
(Compare this exercise with Example 2.5.15.)
36. Let F be a field, and let
2 = {p,(X) : pa(X) is an irreducible monic polynomial in FIXJ}.
We will say that a rational function h(X) = f (X)/g(X) E F(X) is proper
if deg(f (X)) < deg(g(X)). Let F(X)pr denote the set of all proper rational
functions in F[XJ.
(a) Prove that F(X) FIX] tD F(X)pr as F-modules.
(b) Prove that
B=
X' E Z; 0 < j < deg(pa(X)), k > 1 }
(p. (X))'
is a basis of F(X)pr as an F-module. The expansion of a proper rational
function with respect to the basis B is known as the partial fraction
expansion; it should be familiar front elementary calculus.
37. Prove that Q is not a projective Z-module.
38. Let
R = { f : [0, 1] - R : f is continuous and f (0) = f (1) }
and let
M = {f : 10, 11 R: f is continuous and f (0) = - f (1)}.
Then R is a ring under addition and multiplication of functions, and M is
an R-module. Show that M is a projective R-module that is not free. (Hint:
Show that M ®M ? -, R (D R.)
39. Show that submodules of projective modules need not be projective. (Hint:
Consider pZP2 C Z,2 as ZD2-modules.) Over a PID, show that submodules
of projective modules are projective.
40. (a) If R is a Dedekind domain, prove that R is Noetherian.
b If R is an integral domain that is a local ring (i.e., R has a unique
maximal ideal), show that any invertible ideal I of R is principal.
(c) Let R be an integral domain and S C R \ {0} a multiplicatively closed
subset. If I is an invertible ideal of R, show that Is is an invertible ideal
of Rs.
3.9 Exercises 179
(d) Show that in a Dedekind domain R, every nonzero prime ideal is maxi-
mal. (Hint: Let M be a maximal ideal of R containing a prime ideal P,
and let S = R \ M. Apply parts (b) and (c).)
41. Show that Z[f-3] is not a Dedekind domain.
42. Show that Z[XJ is not a Dedekind domain. More generally, let R be any
integral domain that is not a field. Show that RIX] is not a Dedekind domain.
43. Suppose R is a PID and M = R(x) is a cyclic R-module with Ann M = (a) 9
(0). Show that if N is a submodule of M, then N is cyclic with Ann N = (b)
where b is a divisor of a. Conversely, show that M has a unique submodule
N with annihilator (b) for each divisor b of a.
44. Let R be a PID, M an R-module, x E M with Ann(x) = (a) 96 (0). Factor
a = upi' ... pk" with u a unit and pi, ..., pi, distinct primes. Let Y E M
with Ann(y) = (b) 0- (0), where b = ti p1`' . . .pk ' with 0 < rn; < n; for
1 < i < k. Show that Ann(x + y) = (a).
45. Let R be a PID, let M be a free R-module of finite rank, and let N C M be a
submodule. If M/N is a torsion R-module, prove that rank(M) = rank(N).
46. Let R be a PID and let M and N be free R-modules of the same finite rank.
Then an R-module homomorphism f : M - N is an injection if and only if
N/ Im(f) is a torsion R-module.
47. Let u = (a, b) E Z2.
(a) Show that there is a basis of Z2 containing u if and only if a and b are
relatively prime.
(b) Suppose that u = (5,12). Find a v E Z2 such that {u, v} is a basis of
V.
48. Let M be a torsion module over a PID R and assume Ann(M) = (a) 0 (0).
If a = pi' pk'' where pi, ... , pk are the distinct prime factors of a, then
show that MD, = qiM where q; = a/p;'. Recall that if p E R is a prime,
then M9 denotes the p-primary component of M.
49. Let M be a torsion-free R-module over a PID R, and assume that x E M is
a primitive element. If px = qx' show that q I p.
50. Find a basis and the invariant factors for the submodule of Z3 generated by
x1 _ (1,0,-1), x2 = (4,3, -1), x3 = (0,9,3), and x4 = (3, 12,3).
51. Find a basis for the submodule of Q(XJ3 generated by
f, = (2X-1,X,X2+3), f2 = (X,X,X2), f3 = (X + 1, 2X, 2X2 -3).
52. Determine the structure of Z3/K where K is generated by xl = (2,1, -3)

and x2 = (1,-1,2).
53. Let R = RIXJ and suppose that M is a direct sum of cyclic R-modules with
annihilators (X - 1)3, (X2 + 1)2, (X - 1)(X2 + 1)4, and (X + 2)(X2 + 1)2.
Determine the elementary divisors and invariant factors of M.
54. Let R be a PID and let p E R be a prime. Show that submodules, quotient
modules, and direct sums of p-primary modules are p-primary.
55. An R-module M is said to be irreducible if (0) and M are the only sub-
modules of M. Show that a torsion module M over a PID R is irreducible
if and only if M = R(x) where Ann(x) = (p) where p is prime. Show that
if M is finitely generated, then M is indecomposable in the sense that M is
not a direct sum of two nonzero submodules if and only if M = R(x) where
Ann(x) = (0) or Ann(z) = (p°) where p is a prime.
56. Let M be an R-module where R is a PID. We say that M is divisible if for
each nonzero a E R, aM = M.
(a) Show that Q is a divisible Z-module.
(b) Show that any quotient of a divisible R-module is divisible. It follows

for example that Q /Z is a divisible Z-module.
(c) If R is not a field, show that no finitely generated R-module is divisible.
57. Determine all nonisomorphic abelian groups of order 360.
58. Use elementary divisors to describe all abelian groups of order 144 and 168.
59. Use invariant factors to describe all abelian groups of orders 144 and 168.
60. If p and q are distinct primes, use invariant factors to describe all abelian
groups of order
(a) p2g2,
(b) p" q,
(c) P5.
61. If p and q are distinct primes, use elementary divisors to describe all abelian
groups of order p3q2.
62. Let G, H, and K be finitely generated abelian groups. If G x K °_° H x K,

show that G I H. Show by example that this need not be true if we do not
assume that the groups are finitely generated.
63. Determine all integers for which there exists a unique abelian group of order
n.
64. Show that two finite abelian groups are isomorphic if and only if they have
the same number of elements of each order.
65. Let p be a prime and assume that a finite abelian group G has exactly k
elements of order p. Find all possible values of k
66. Find a generator for the cyclic group F' where F is each of the following
fields (see Example 2.5.15 (3)):
(a) F2(XJ/(X2 + X + 1).
(b) F3[X]/(X2 + 1).
67. Let
0 M1
-
M2
/s+ ... "- Mn+l -+ 0
be an exact sequence of finite rank free modules and homomorphisms over a
PID R. That is, fl is injective, fn is surjective, and Im(f1) = Ker(f;+l) for
1 <i<n-1. Show that
n+l
(-1)'}1 rank(M.) = 0.
68. If f(XI, ... Xn) E R[X1, ... X 1 , , the degree off is the highest degree of
a monomial in f with nonzero coe cient, where
deg(X" ...X'") = it +...+in.

Let F be a field. Given any five points {v1, ... ,v5} C F2, show that there
is a quadratic polynomial f (Xl, X2) E F[X1, X21 such that f (v;) = 0 for
1<i<5.
69. Let M and N be finite-rank free R-modules over a PID R and let f E
HomR(M, N). If S C N is a complemented submodule of N, show that
f '(S) is a complemented submodule of M.
70. Let R be a PID, and let f : M -. N be an R-module homomorphism of
finite rank free R-modules. If S C N is a submodule, prove that
rank(f -1(S)) = rank(S fl Im(f )) + rank(Ker(f)).
3.9 Exercises 181
71. Let M1 f M L M2 be a sequence of finite-rank R-modules and R-

module homomorphisms, where R is a PID, and assume that Im(f) is a
complemented submodule of M.
(a) Show that
rank(Im(g o f)) = rank(lm(f)) - rank(Im(f) fl Ker(g)).
(b) Show that
rank(lm(g o f)) = rank(Im(f) + Ker(g)) - rank(Ker(g)).
If R is a field, then all submodules of R-modules are complemented, so
these formulas are always valid in the case of vector spaces and linear
transformations. Show, by example, that they need not be valid if Im(f)
is not complemented.
72. Let R be a PID, and let M, N, and P be finite rank free R -modules. Let
M - N and g : M - P be homomorphisms. Suppose that Ker(f) C Ker(g)
and Im(f) is a complemented submodule of N. Then show that there is a
homomorphism h: N -+ P such that g = h o f.
73. Let F be a field and let V be a vector space over F. Suppose that f, g E
V' = HomF(V, F) such that Ker(f) C Ker(g). Show that there is a E F
such that g = a f . this same result true if F is replaced by a PID?
74. Let R be a PID and let M be a finite rank free R-module. Let Ck(M) denote
the set of complemented submodules of M of rank k. Let G be the group of
units of the ri EndR(M).
(a) Show that-710, N) -. b(N) determines an action of the group C on the
set Ck(M).
(b) Show that the action defined in part (a) is transitive, i.e., given N1,
N2 E Ck(M) there is 0 E G that sends N1 to N2.
I. Modules
The algebraic categories with which we shall be principally concerned

in this book are categories of modules over a fixed (unitary) ring A and
module-homomorphisms. Thus we devote this chapter to a preliminary
discussion of A-modules.
The notion of A-module may be regarded as providing a common
generalization of the notions of vector space and abelian group. Thus
if A is a field K then a K-module is simply a vector space over K and a
K-module homomorphism is a linear transformation; while if A = 7l
then a 7l-module is simply an abelian group and a 7l-module homo-
morphism is a homomorphism of abelian groups. However, the facets
of module theory which are of interest in homological algebra tend to be
trivial in vector space theory; whereas the case A = 7l will often yield
interesting specializations of our results, or motivations for our construc-
tions.
Thus, for example, in the theory of vector spaces, there is no interest
in the following question: given vector spaces A, B over the field K,
find all vector spaces E over K having B as subspace with A as associated
quotient space. For any such E is isomorphic to AEeB. However, the
question is interesting if A, B, E are now abelian groups; and it turns
out to be a very basic question in homological algebra (see Chapter III).
Again it is trivial that, given a diagram of linear transformations of
K-vector spaces
(0.1)
B~C
where e is surjective, there is a linear transformation {3: P->B with

e{3 = y. However, it is a very special feature of an abelian group P that,
for all diagrams of the form (0.1) of abelian groups and homomorphisms,
with e surjective, such a homomorphism {3 exists. Indeed, for abelian
groups, this characterizes the free abelian groups (thus one might say
that all vector spaces are free). Actually, in this case, the example A = 7l
is somewhat misleading. For if we define a A-module P to be projective if,
given any diagram (0.1) with e surjective, we may find {3 with e{3=y,
1. Modules 11
then it is always the case that free A-modules are projective but, for some
rings A, there are projective A-modules which are not free . The relation
between those two concepts is elucidated in Sections 4 and 5, where we
see that the concepts coincide if A is a principal ideal domain (pj.d.) -
this explains the phenomenon in the case of abelian groups.
In fact, the matters of concern in homological algebra tend very much
to become simplified - but not trivial - if A is a pj.d., so that this special
case recurs frequently in the text. It is thus an important special case, but
nevertheless atypical in certain respects. In fact, there is a precise numerical
index (the so-called global dimension of A) whereby the case A a field
appears as case 0 and A a pj.d. as case 1.
The categorical notion of duality (see Chapter II) may be applied to
the study of A-modules and leads to the concept of an injective module,
dual to that of a projective module. In this case, the theory for A = 7L,
or, indeed, for A any pj.d., is surely not as familiar as that of free modules ;
nevertheless, it is again the case that the theory is, for modules over a pj.d.,
much simpler than for general rings A - and it is again trivial for vector
spaces!
We should repeat (from the main Introduction) our rationale for
placing this preparatory chapter on modules before the chapter introduc-
ing the basic categorical concepts which will be used throughout the
rest ofthe book. Our justification is that we wish, in Chapter II, to have
some mathematics available from which we may make meaningful
abstractions. This chapter provides that mathematics; had we reversed
the order of these chapters, the reader would have been faced with a
battery of "abstract" ideas lacking in motivation. Although it is, of course,
true that motivation, or at least exemplification, could in many cases
be provided by concepts drawn from other parts of mathematics familiar
to the reader, we prefer that the motivation come from concrete instances
of the abstract ideas germane to homological algebra.
1. Modules
We start with some introductory remarks on the notion of a ring. In

this book a ring A will always have a unity element lA =l= O. A homo-
morphism of rings w: A ---+ r will always carry the unity element of the
first ring A into the unity element of the second ring r. Recall that the
endomorphisms of an abelian group A form a ring End(A, A).
Definition. A left module over the ring A or a left A-module is an
abelian group A together with a ring homomorphism w : A ---+ End (A, A).
We write Aa for (W(A)) (a), a E A, AE A. We may then talk of A operating
(on the left) on A, in the sense that we associate with the pair (A, a) the
12 I. Modules
element Aa. Clearly the following rules are satisfied for all a, al ' a2 E A,
A, Al' A2 E A:
M 1: (Al + A2)a = Al a + A2 a
M 2: (Al A2)a = Al (A2 a)
M3: l A a=a
M4: A(a l +a 2)=Aa l +Aa 2 ·
On the other hand, if an operation of A on the abelian group A
satisfies M 1, .. . , M 4, then it obviously defines a ring homomorphism
w :.A - End(A, A), by the rule (W(A)} (a) = Aa .
Denote by AOPP the opposite ring of A. The elements AOPP E AOPP are
in one-to-one correspondence with the elements A E A. As abelian groups
A and AOPP are isomorphic under this correspondence. The product in
AOPP is given by Ai PP A~PP = (A2 Al)OPP. We naturally identify the underlying
sets of A and AOPP.
A right module over A or right A-module is simply a left AOPP-module,
that is, an abelian group A together with a ring map w' : AOPP_ End(A, A).
We leave it to the reader to state the axioms M 1', M 2', M 3', M 4' for a
right module over A. Clearly, if A is commutative, the notions of a left
and a right module over A coincide. For convenience, we shall use the
term "module" always to mean "left module".
Let us give a few examples:
(a) The left-multiplication in A defines an operation of A on the
underlying abelian group of A, satisfying M 1, .. . , M 4. Thus A is a left
module over A . Similarly, using right multiplication, A is a right module
over A. Analogously, any left-ideal of A becomes a left module over A,
any right-ideal of A becomes a right module over A.
(b) Let A = 7l, the ring of integers. Every abelian group A possesses
the structure of a 7l-module; for a E A , nE 7l define n a = 0, if n = 0,
na=a+ ... +a (n times), ifn>O, and na= -(-na), ifn<O.
(c) Let A = K, a field . A K-module is a vector space over K .
(d) Let V be a vector space over the field K, and T a linear trans-
formation from V into V. Let A = K [T], the polynomial ring in T
over K. Then V becomes a K[T]-module, with the obvious operation
of K[T] on V.
(e) Let G be a group and let K be a field . Consider the K-vector-
space of all linear combinations L kxx , kxE K. One checks quite
xeG
easily that the definition
where xy denotes the product in G, makes this vector space into a K-

algebra KG, called the group algebra of Gover K. Let V be a vector space
1. Modules 13
over K. A K-representation of G in V is a group homomorphism

a: G---+ AutK(V, V). The map a gives rise to a ring homomorphism
a' : K G---+ EndK(V, V) by setting
(J' ( L kxx) = L kx(J(x).

XEG XEG
Since every K -linear endomorphism of V is also a homomorphism of the

underlying abelian group, we obtain from (J' a ring homomorphism
(!: KG---+ End z (V, V), making V into a KG-module. Conversely, let V
be a KG-module. Clearly V has a K-vector-space structure, and the struc-
ture map (! : K G---+ Endz(V, V) factors through EndK(V, V). Its restriction
to the elements of G defines a K-representation of G. We see that the
K-representations of G are in one-to-one correspondence with the KG-
modules. (We leave to the reader to check the assertions in this example.)
Definition . Let A, B two A-modules. A homomorphism (or map)
<p: A---+B of A-modules is a homomorphism of abelian groups such that
<p(Aa) = A(<pa) for all a E A, A E A.
Clearly the identity map of A isa homomorphism of A-modules;
we denote it by 1A : A---+A.
If ---->B. We call <p : A---+ B isomorphic or an isomorphism,
and write <p : A~ B, if there exists a homomorphism lp : B---+ A such that
lp<p = 1A and <plp = lB' Plainly, if it exists, lp is uniquely determined;
it is denoted by <p - 1 and called the inverse of <po If <p : A ---+ B is isomorphic,
it is clearly injective and surjective. Conversely, if the module homo-
morphism <p: A ---+ B is both injective and surjective, it is isomorphic.
We shall call A and B isomorphic, A ~ B, if there exists an isomorphism
<p: A~B.
If A' is a subgroup of A with Ad E A' for all AE A and all a' E A',
then A' together with the induced operation of A is called a submodule
of A. Let A' be a submodule of A. Then the quotient group A/ A' may be
given the structure of a A-module by defining A(a + A') = (Aa + A')
for aliA E A, a EA. Clearly, we have an injective homomorphisllj1: A'>---->A
and a surjective homomorphism n: A-A/ A'.
F or an arbitrary homomorphism <p: A ---+ B, we shall use the nota-
tion ker <p = {a E A I <pa = O} for the kernel of <p and
im <p = <p A = {b E Bib = <p a for some a E A}
for the image of <po Obviously ker <p is a submodule of A and im <p is
a submodule of B. One easily checks that the canonical isomorphism
of abelian groups A/ker cp':"'im cp is actually an isomorphism of A-modules.
We also introduce the notation coker ---->A is injective,
14 I. Modules
we can identify A' with the submodule J-LA ' of A. Similarly, if t:: A-A"
is surjective, we can identify A" with A/ker t:.
Definition. Let <{J: A-B and 1p: B-C be homomorphisms of A-
modules. The sequence A~B~C is called exact (at B) ifker1p=im<{J.
If a sequence Ao-AI- ···-An-An+1 is exact at AI' ... , An, then the
sequence is simply called exact.
As examples we mention
(a) O-A~B is exact (at A) if and only if <p is injective.
(b) A~B-O is exact (at B) if and only if <p is surjective.
(c) The sequence O_A'-4A~A"_O is exact (at A', A, A") if and
only if J-L induces an isomorphism A' -=+ J-L A' and t: induces an isomorphism
A/kert: = A/J-LA'-=+A". Essentially A' is then a submodule of A and A" the
corresponding quotient module. Such an exact sequence is called short
exact, and often written A'>--->A-A".
The proofs of these assertions are left to the reader. Let A, B, C, D
be A-modules and let IX, (3, )', 6 be A-module homomorphisms. We say
that the diagram
A~B
1 1
y p
C---L.D
is commutative if (31X = 6)': A-D. This notion generalizes in an obvious
way to more complicated diagrams. Among the many propositions and
lemmas about diagrams we shall need the following:
Lemma 1.1. Let A'>--->A-A" and B'>--->B-B" be two short exact
sequences. Suppose that in the commutative diagram
(1.2)
any two of the three homomorphisms IX', IX, IX" are isomorphisms. Then the
third is an isomorphism, too.
Proof. We only prove one of the possible three cases, leaving the
other two as exercises. Suppose IX', IX" are isomorphisms; we have to
show that IX is an isomorphism.
First we show that ker IX = O. Let a E kerrx, then 0 = t:' IX a = IX" w.
Since IX" is an isomorphism, it follows that w = O. Hence there exists
dEA' with J-Ld = a by the exactness of the upper sequence. Then
0= IXJ-Ld = J-L'IX' d. Since J-L' IX' is injective, it follows that d = O. Hence
a = J-Ld =0.
1. Modules 15
Secondly, we show that a is surjective. Let bE B; we have to show

that b = a a for some a E A. Since a" is an isomorphism, there exists
a" E A" with a" a" = e' b. Since e is surjective, there exists Ci E A such
that eCi = a". We obtain e'(b -ali) = e' b - e' aCi = e' b -a" eCi = O. Hence
by the exactness of the lower sequence there exists b' E B' withJ-L' b' = b -aCi.
Since a' is isomorphic there exists a' E A' such that a' a' = b'. Now
a(J-La' + li) = a J-L a' + aCi = J-L' a' a' + aCi = J-L' b' + aCi = b .
So setting a=J-La' +a, we have aa=b. 0
Notice that Lemma 1.1 does not imply that, given exact sequences
A',........ A-A", B',........B-B", with A'~B', A"~B", then A~B. It is
crucial to the proof of Lemma 1.1 that there is a map A-B compatible
with the isomorphisms A' ~ B', A" ~ B", in the sense that (1.2) commutes.
Exercises:
1.1. Complete the proof of Lemma 1.1. Show moreover that, in (l.2), IX is surjective
(injective) if IX', IX" are surjective (injective).
1.2. (Five Lemma) Show that, given a commutative diagram
---+ B I ---+ B 2 ---+ B 3 ---+ B4 ---+ B s ---+
with exact rows, in which Ipl, 1p2, 1p4, Ips are isomorphisms, then 1p3 is also an
isomorphism. Can we weaken the hypotheses in a reasonable way?
1.3. Give examples of short exact sequences of abelian groups
0--> A' --> A --> A" -->0, 0--> B' --> B--> B"-->O
such that
(i) A'~B' , A~B, A"*B";
(ii) A'~B', A*B, A"~B";
(iii) A'*B', A~B , A"~B" .
tA. Show that the abelian group A admits the structure of a Zm-module if and
only if mA = O.
1.5. Define the group algebra KG for K an arbitrary commutative ring. What are
the KG-modules?
1.6. Let V be a non-trivial (left) KG-module. Show how to give V the structure of
a non-trivial right KG-module. (Use the group inverse.)
1.7. Let 0-->A'4A~A"-->0 be a short exact sequence of abelian groups. We say
that the sequence is pure if, whenever f/(a') = ma, a' E A', aE A, m a positive
integer, there exists b' E A' with a' = mb'. Show that the following statements
are equivalent:
(i) the sequence is pure;
16 1. Modules
(ii) the induced sequence (reduction mod m) 0--> A~~ Am~ A~--> 0 is
exact for all m; (Am = A/mA, etc.)
(iii) given a" EA" with ma" =O, there exists aEA with e(a) = a", ma=O
(for all m).
2. The Group of Homomorphisms
Let HomA(A, B) denote the set of all A-module homomorphisms from

A to B. Clearly, this set has the structure of an abelian group ; if cp : A --> B
and tp : A-->B are A-module homomorphisms, then cp + tp : A-->B is
defined as (cp + tp)a = cpa + tpa for all a E A. The reader should check
thatcp + tpisaA-modulehomomorphism. Note, however, that HomA(A ,B)
is not, in general, a A-module in any obvious way (see Exercise 2.3).
Let [3 : BI-->B2 be a homomorphism of A-modules. We can assign
to a homomorphism cp: A --> B I , the homomorphism [3 cp : A --> B 2 , thus
defining a map [3* = HomA(A, [3): HomA(A, BI )--> HomA(A , B2)' It is left
to the reader to verify that [3* is actually a homomorphism of abelian
groups. Evidently the following two rules hold:
(i) If [3 : BI--> B2 and [3' : B 2--> B 3 , then
([3' [3)* = [3~ [3* : HomA(A, B I )--> Hom A(A , B3 ) .
(ii) If [3 : BI--> BI is the identity, then [3* : HomA(A , BI )--> Hom A(A, B I )
is the identity, also.
In short, the symbol HomA(A , -) assigns to every A-module B an
abelian group HomA(A, B), and to every homomorphism of A-modules
[3 : BI-->B2 a homomorphism of abelian groups
[3* = HomA(A , [3): HomA(A , BI )--> HomA(A , B2)
such that the above two rules hold. In Chapter II, we shall see that this
means that HomA(A, -) is a (covariant) functor from the category of
A-modules to the category of abelian groups.
On the other hand, if 0(: A2-->AI is a A-module homomorphism,
then we assign to every homomorphism cp : Al --> B the homomorphism
CPO(: A2-->B, thus defining a map
0(* = Hom:.t(O(, B) : HomA(A I , B)-->HomA(A 2, B).
Again we leave it to the reader to verify that 0(* is actually a homomorphism

of abelian groups. Evidently, we have:
(i)' If 0( : A 2--> Al and 0(' : A 3--> A 2, then (0(0(')* = 0('* 0(* (inverse order !).
(ii)' If 0( : AI--> Al is the identity, then 0(* is the identity.
HomA( - , B) is an instance ofa contravariant functor (from A-modules
to abelian groups).
2. The Group of Homomorphisms 17
Theorem 2.1. Let B',!!.....B~B" be an exact sequence of A-modules.

For every A-module A the induced sequence
O~HomA(A, B')~HomA(A, B)~HomA(A, B")

is exact.
Proof· First we show that 11* is injective.

Assume that I1<P in the diagram
A
1~
is the zero map. Since 11: B',........,.B is injective this implies that <p : A~B'
is the zero map, so 11* is injective.
Next we show that ker 8* ) im 11*. Consider the above diagram.
A map in im 11* is of the form 11 <po Plainly 811 <p is the zero map, since 811
already is. Finally we show that iml1*) kerB*. Consider the diagram
A
lw
B'~B~B".
We have to show that if 81p is the zero map, then 1p is of the form 11<P
for some <P : A~B' . But,if81p = Otheimage of1piscontainedinker8 = iml1.
Since 11 is injective, 1p gives rise to a (unique) map <P: A~B' such that
11<P = 1p. 0
We remark that even in case 8 is surjective the induced map 8* is not
surjective in general (see Exercise 2.1).
Theorem 2.2. Let A ' -4A~A " be an exact sequence of A-modules.
For every A-module B the induced sequence
O~ HomA (A", B)~ HomA(A, B)4 Hom A (A' , B)

is exact.
The proof is left to the reader. 0
Notice that even in case 11 is injective 11* is not s"urjective in general
(see Exercise 2.2).
We finally remark that Theorem 2.1 provides a universal characteri-
zation of kerB (in the sense of Sections 11.5 and 11.6): To every homo-
morphism <p : A~B with 8 *(<p)= 8 <p:A~B" the zero map there exists
a unique homomorphism <p': A~B' with 11*(<P') = 11<P' = <po Similarly
Theorem 2.2 provides a universal characterization of coker 11.
18 I. Modules
Exercises:
2.1. Show that in the setting of Theorem 2.1 e* = Hom(A, e) is not, in general,
surjective even if e is. (Take A = 7l, A = 7l n , the integers mod n, and the short
exact sequence 7l'!!""'7l-7ln where 11 is multiplication by n.)
2.2. Prove Theorem 2.2. Show that 11* = Hom A(II, B) is not, in general, surjective
even if 11 is injective. (Take A = 7l, B = 7l n , the integers mod n, and the short
exact sequence 7l'!!""'7l-7l n , where 11 is multiplication by n.)
2.3. Suppose A commutative, and A and B two A-modules. Define for a A-module
homomorphism ({J: A -+ B, (A({J) (a) = ((J(Aa), a EA. Show that this definition
makes HomA(A, B) into a A-module. Also show that this definition does not
work in case A is not commutative.
2.4. Let A be a A-module and B be an abelian group . Show how to give
Homz(A, B) the structure of a right A-module.
2.5. Interpret and prove the assertions 0* = 0, 0* = O.
2.6. Compute Hom(7l,7l n ), Hom(7lm, 7l n ), Hom(7lm,71:), Hom(<Q,7l), Hom(<Q, <Q).
[Here "Hom" means "Homz" and <Q is the group of rationals.]
2.7. Show (see Exercise 1.7) that the sequence O-+A'-+A-+A"-+O is pure if and
only if Hom(7lm, -) preserves exactness, for all m > O.
2 .S. If A is a left A-module and a right r-module such that the A-action commutes
with the r-action, then A is called a left A-right r-bimodule. Show that if A
is a left A-right l"-bimodule and B is a left A-right r-bimodule then HomA(A, B)
is naturally a left l"-right r-bimodule.
3. Sums and Products
Let A and B be A-modules. We construct the direct sum A ey B of A and B

as the set of pairs (a, b) with a E A and bE B together with componentwise
addition (a, b) + (a', b' ) = (a + a', b + b') and componentwise A-operation
A(a, b) = (Aa, Ab). Clearly, we have injective homomorphisms ofA-modules
'A:A-,>-AE8B defined by IA(a)=(a, 0) and 'B: B-'>-AE8B defined by IB(b) =
(0, b).
Proposition 3.1. Let M be a A-module, tpA:A--+M and tpB:B--+M
A-module homomorphisms. Then there exists a unique map
tp = <1JJA' tpB): AeyB--+M

such that tp I A = 1JJ A and tp 'B = 1JJB'
We can express Proposition 3.1 in the following way: For any A-
module M and any maps 1JJA' 1JJB the diagram
3. Sums and Products 19
can be completed by a unique homomorphism 1p: AE£lB---->M such

that the two triangles are commutative.
In situations like this where the existence of a map is claimed which
makes a diagram commutative, we shall use a dotted arrow to denote
this map. Thus the above assertion will be summarized by the diagram
A~
~M
B
and the remark that 1p is uniquely determined.
Proof. Define 1p(a, b) = 1p A(a) + 1pB(b). This obviously is the only
homomorphism 1p: A EB B ----> M satisfying 1p I A = 1p A and 1p I B = 1p B ' D
We can easily expand this construction to more than two modules:
Let {A j }, j E J be a family of A-modules indexed by J. We define the
direct sum EBAj of the modules Aj as follows: An element of AjEB
jEJ j EJ
is a family (a)jd with aj E Aj and aj =1= 0 for only a finite number of sub-
scripts. The addition is defined by (a)jEJ + (b)jEJ = (aj + b)jEJ and the
A-operation by A(a)jEJ = (Aa)jEJ' For each k E J we can define injections
EB
I k : Ak----> Aj by lk(a k ) = (bj)jEJ with bj = 0 for j =1= k and bk = a b ak E A k •
jEJ
Proposition 3.2. Let M be a A-module and let {1pi: A;----> M}, j E J,

be a family of A-module homomorphisms. Then there exists a unique homo-
morphism 1p = <1pj >:EB
Aj----> M, such that 1p Ij = 1pj for all j E J.
jEJ
Proof. We define 1p((a)jEJ) = I 1p/a). This is possible because a/ = 0

jEJ
except for a finite number of indices. The map 1p so defined is obviously
the only homomorphism 1p : EB
Aj----> M such that 1plj = 1pj for allj E J. D
jEJ
We remark the important fact that the property stated in Pro-
position 3.2 characterizes the direct sum together with the injections up to
a unique isomorphism. To see this, let the A-module S together with
injections Ij: Aj---->S also have the property f!J> claimed for ($
JEJ
Aj; Ij)
in Proposition 3.2. Write (temporarily) T for
jEJ
EB
Aj . First choose M = T
and 1pj = lj' j E J. Since (S; Ij) has property f!J>, there exists a unique
homomorphism 1p: S----> T such that the diagram
A.~
JI~ ~
Ij
S········ ······~T
v'
20 I. Modules
is commutative for every j E J. Choosing M = Sand 1pj = Ij and invoking

property f!1> for (T; Ij) we obtain a map 1p' : T --> S such that the diagram
is commutative for every j E J. In order to show that 1p1p' is the identity,

we remark that the diagram
A'~lj
\~ T::::::::::::::::,)T
lPlP'
is commutative for both 1p1p' and the identity. By the uniqueness part of
property f!1> we conclude that 1p1p' = IT' Similarly we prove that 1p'1p = Is·
Thus both 1p and 1p' are isomorphisms.
A property like the one stated in Proposition 3.2 for the direct sum
of modules is called universal. We shall treat these universal properties
in detail in Chapter II. Here we are content to remark that the construction
ofthe direct sum yields an existence proof for a module having property f!1>.
Next we define the direct product TI
Aj ofa family of modules {AJ,jEJ.
jeJ
An element of TI Aj is a family (a)jeJ of elements aj E Aj. No restrictions
jeJ
are placed on the elements aj ; in particular, the elements aj may be non-
zero for an infinite number of subscripts. The addition is defined by
(a)jeJ + (b)jeJ = (aj + b)jeJ and the A-operation by A(a)jeJ = (Aaj)jeJ '
TI
For each k E J we can define projections 1T. k : Aj-->Ak by 1T. k(aj)jeJ = ak.
jeJ
For a finite family of modules Aj,j = 1, ... , n, it is readily seen that the
n n
modules TI Aj and EB Aj are identical; however in considering the

j=l j=l
direct sum we put emphasis on the injections I j and in considering the
direct product we put emphasis on the projections 1T.j .
Proposition 3.3. Let M be a A -module and let {C{Jj: M --> A j}, j E J,
be a family of A-module homomorphisms. Then there exists a unique homo-
morphism C{J = {C{Jj} : !VI --> TI Aj such that for every j E J the diagram
jeJ
~/
TI
M HH' Aj
'P jeJ
is commutative, i.e. 1T.jC{J = C{Jj. D
3. Sums and Products 21
The proof is left to the reader; also the reader will see that the universal
property of the direct product TI
Aj and the projections nj characterizes
j EJ
it up to a unique isomorphism. Finally we prove
Proposition 3.4. Let B be a A-module and {Aj},j E J be a family of A-
modules. Then there is an isomorphism
I] : HomA (EB A
jEJ
j , B) ...::. TI HomA(Aj, B) .
jEJ
Proof. The proof reveals that this theorem is merely a restatement of

the universal property of the direct sum. For 1p: Ar..... B, define EB
jEJ
(1p lj : A r . . . B)jEJ' Conversely a family {1pj : Aj-> B}, j E J, gives rise
I](1p) =
to a unique map 1p: EB Aj->B. The projections nj: HomA(Aj, B) TI
jEJ jEJ
->HomA(Aj, B) are given by njl] = HomA(lj, B). 0
Analogously one proves:
Proposition 3.5. Let A be a A-module and {Bj }, j E J be a family of
A-modules. Then there is an isomorphism
(: HomA (A, TI B j )"'::' TI HomA(A, B).
jEJ jEJ
The proof is left to the reader. 0

Exercises:
3.1. Show that there is a canonical map a: EB Ar-> n A j •
m
3.2. Show how a map from EB Ai to EB B may be represented by a matrix
j
i=l j= 1
where rpij: Ai->Bj . Show that, if we write the composite of rp :A->B and
lP: B->C as rptp (not tprp), then the composite of
m
I[J = (rpij): EB Ai-> EB B j
i= l j = l
and
n q
'P = (lPjk): EB Bj-> EB Ck
j ~ l k ~ l
is the matrix product I[J 'P.

3.3. Show that if, in (1.2), (1.' is an isomorphism, then the sequence
O->A~A"EBB (o"' - '">,B"->O
is exact. State and prove the converse.

22 I. Modules
3.4. Carry out a similar exercise to the one above, assuming rx" is an isomorphism.
3.5. Use the universal property of the direct sum to show that
(AI EBA 2 )EBA 3 ~ Al EB(A 2 EBA 3 )·
3.6. Show that 7l m EB71 n =71 mn if and only if m and n are mutually prime.
3.7. Show that the following statements about the exact sequence
O->A'-4A!4A" ->O
of A-modules are equivalent:

(i) there exists /1 : A" -> A with rx" /1 = I on An;
(ii) there exists e: A -> A' with erx' = 1 on A';
(iii) 0-> HomA(B, A')~HomA(B, A)~HomA(B, A")->O is exact for all B;
(iv) 0-> HomA(A", q~HomA(A, q~HomA(A', q->O is exact for all C;
(v) there exists /1: A"-> A such that <rx', /1) : A' EB An -='A.
3.8. Show that if O->A'-4A!4A"->O is pure and if A" is a direct sum of cyclic
groups then statement (i) above holds (see Exercise 2.7).
4. Free and Projective Modules
Let A be a A-module and let S be a subset of A. We consider the set Ao

of all elements a E A of the form a = L
As S where As E A and As =F 0 for
seS
only a finite number of elements s E S. It is trivially seen that Ao is a
submodule of A; hence it is the smallest submodule of A containing S.
If for the set S the submodule Ao is the whole of A, we shall say that S
is a set of generators of A. If A admits a finite set of generators it is said
to be finitely generated. A set S of generators of A is called a basis of A
if every element a E A may be expressed uniquely in the form a = LAss
seS
with As E A and As =F 0 for only a finite number of elements s E S. It is
readily seen that a set S of generators is a basis if and only if it is linearly
independent, that is, if LAss = 0 implies As = 0 for all S E S. The reader
seS
should note that not every module possesses a basis.
Definition. If S is a basis of the A-module P, then P is called free on the
set S. We shall call P free ifit is free on some subset.
Proposition 4.1. Suppose the A-module P is free on the set S. Then

P~ EB A s where As = A as a left module for S E S. Conversely, As EB
seS seS
is free on the set {lA" S E S}.
Proof. We define cp: P- EB As as follows: Every element a E P is

seS
expressed uniquely in the form a= 2: AsS; set cp(a) = (As}ses . Conversely,
SES
4. Free and Projective Modules 23
for s E S define ips: As-P by ips (A,) = AsS' By the universal property ofthe
<
direct sum the family {ips}, S E S, gives rise to a map ip = ips> : EB
As- P.
SES
It is readily seen that cp and ip are inverse to each other. The remaining
assertion immediately follows from the construction of the direct sum. 0
The next proposition yields a universal characterization of the free
module on the set S.
Proposition 4.2. Let P be free on the set S. To every A-module M and
to every function f from S into the set underlying M, there is a unique
A-module homomorphism cp: P-M extending f.
Proof. Let f(s) = ms' Set cp(a) = cp (I AsS) = I
SES ) SES
Asms. This obviously
is the only homomorphism having the required property. 0
Proposition 4.3. Every A-module A is a quotient of a free module P.
Proof. Let S be a set of generators of A. Let P = EB As with As = A
SES
and define cp: P-A to be the extension of the function f given by

f(lA) = s. Trivially cp is surjective. 0
Proposition 4.4. Let P be a free A-module. To every surjective homo-
morphism e: B-C of A-modules and to every homomorphism y: P-C
there exists a homomorphism f3 : P - B such that e f3 = y.
Proof. Let P be free on S. Since e is surjective we can find elements
bs E B, s E S with e(b s ) = y(s), s E S. Define f3 as the extension of the func-
tion f: S - B given by f(s) = bs' s E S. By the uniqueness part of Pro-
position 4.2 we conclude that ef3 = y. 0
To emphasize the importance of the property proved in Proposition 4.4
we make the following remark: Let A A B~ C be a short exact sequence
of A-modules. If P is a free A-module Proposition 4.4 asserts that every
homomorphism y: P-C is induced by a homomorphism f3: P-B.
Hence using Theorem 2.1 we can conclude that the induced sequence
is exact, i.e. that e* is surjective. Conversely, it is readily seen that exactness

of (4.1) for all short exact sequences A>->B- C implies for the module
P the property asserted in Proposition 4.4 for P a free module. Therefore
there is considerable interest in the class of modules having this property.
These are by definition the projective modules:
Definition. A A-module P is projective if to every surjective homo-
morphism e: B-C of A-modules and to every homomorphism y: P-C
there exists a homomorphism f3 : P - B with e f3 = y. Equivalently, to any
homomorphisms e, y with e surjective in the diagram below there exists
1. Modules
24
{3 such that the triangle

p
k,••.•.• ........
B - - -...
"C
is commutative.
As mentioned above, every free module is projective. We shall give
some more examples of projective modules at the end of this section.
Proposition 4.5. A direct sum E8 Pi is projective if and only if each Pi is.
;eI
Proof. We prove the proposition only for A = P ffi Q. The proof in the
general case is analogous. First assume P and Q projective. Let c: B-C
be surjective and y: P EB Q~ C a homomorphism. Define yp = Ylp: P~C
and YQ = YzQ : Q~ C. Since P, Q are projective there exist {3p, {3Q such that
a{3p = YP' a{3Q = YQ. By the universal property of the direct sum there
exists {3:PEBQ~B such that {3lp={3p and {3zQ={3Q. It follows that
(a{3) Zp = a{3p = yp = YIp and (a{3) lQ = a{3Q = YQ = YlQ. By the uniqueness
part of the universal property we conclude that a{3 = y. Of course, this
could be proved using the explicit construction of P ffi Q, but we prefer
to emphasize the universal property of the direct sum.
Next assume that PEBQ is projective. Let a: B-C be a surjection
and yp: P~C a homomorphism. Choose YQ: Q~C to be the zero map.
We obtain y: PEBQ~C such that yZp = yp and YZQ = YQ = O. Since PEBQ
is projective there exists {3 : P EB Q~B such that a{3 = y. Finally we obtain
a({3zp) = yip = yp. Hence {3lp: P~B is the desired homomorphism. Thus P
is projective; similarly Q is projective. 0
In Theorem 4.7 below we shall give a number of different characteriza-
tions of projective modules. As a preparation we define :
Definition . A short exact sequence A4B~C of A-modules splits if
there exists a right inverse to e, i.e. a homomorphism a: C---+B such that
£a= Ie. The map a is then called a splitting.
We remark that the sequence A~AEBC~C is exact, and splits
by the homomorphism lc. The following lemma shows that all split short
exact sequences of modules are of this form (see Exercise 3.7).
Lemma 4.6. Suppose that a: C~ B is a splitting for the short exact
sequence A~B~C. Then B is isomorphic to the direct sum AEBC.
Under this isomorphism, J1 corresponds to IA and a to I C .
In this case we shall say that C (like A) is a direct summand in B.
Proof. By the universal property of the direct sum we define a map 1p
as follows
A~
~B
4. Free and Projective Modules 25
Then the diagram
A~B '"C
is commutative; the left-hand square trivially is; the right-hand square
is by et/!(a, c)=e(I.la+uc)=O+euc=c, and 1tc(a, c)=c, a E A, c E C. By
Lemma 1.1 t/! is an isomorphism. 0
Theorem 4.7. For a A -module P the following statements are equivalent:
(1) P is projective;
(2) for every short exact sequence AhB~C of A-modules the
induced sequence
is exact;
(3) if e : B- P is surjective, then there exists a homomorphism f3 : P---+ B
such that e f3 = 1p;
(4) P is a direct summand in every module of which it is a quotient;
(5) P is a direct summand in a free module.
Proof. (1)=>(2). By Theorem 2.1 we only have to show exactness at
HomA(P, C), i.e. that e* is surjective. But since e: B---+ C is surjective this
is asserted by the fact that P is projective.
(2)=>(3). Choose as exact sequence kere>---+B-4.P. The induced
sequence
0---+ HomA (P, ker e)---+ HomA (P, B)!!.4 HomA(P, P)---+O
is exact. Therefore there exists f3 : P---+ B such that e f3 = 1p.
(3)=>(4). Let P ~ BjA, then we have an exact sequence A>---+B-4P.
By (3) there exists f3: P---+B such that ef3 = Ip. By Lemma 4.6 we conclude
that P is a direct summand in B.
(4)=>(5). By Proposition 4.3 P is a quotient of a free module P'.
By (4) P is a direct summand in P'.
(5)=>(1). By (5) P'~PtBQ, where P' is a free module. Since free
modules are projective, it follows from Proposition 4.5 that P is
projective. 0
Next we list some examples:
(a) If A = K, a field, then every K-module is free, hence projective.
(b) By Exercise 2.1 and (2) of Theorem 4.7, tl n is not projective as a
module over the integers. Hence a finitely generated abelian group is
projective if and only if it is free.
(c) Let A = 7l 6, the ring of integers modulo 6. Since 716 = 713 tB712
as a 7l 6 -module, Proposition 4.5 shows that 712 as well as 713 are projective
7l 6 -modules. However, they are plainly not free 7l 6 -modules.
1. Modules
26
Exercises:
4.1. Let V be a vector space of countable dimension over the field K. Let
A = HomK(V, V). Show that, as K-vector spaces V, is isomorphic to VEB v.
We therefore obtain
A=HomK(V, V)~HomK(VEBV, V)~HomK(V, V)EBHomK(V, V)=AEBA.
Conclude that, in general, the free module on a set of n elements may be iso-
morphic to the free module on a set of m elements, with n ~ m.
4.2. Given two projective A-modules P, Q, show that there exists a free A-module R
such that PEBR ~ QEBR is free. (Hint: Let PEBP' and QEBQ' be free. Define
R=FEB~EBmEBWEBnEB···~gEBWEBnEB~EBmEB···J
4.3. Show that <Q is not a free 'Z-module.
4.4. Need a direct product of projective modules be projective?
4.5. Show that if O->N->P->A->O, O->M->Q->A->O are exact with P, Q
projective, then P EB M ~ Q EB N. (Hint: Use Exercise 3.4.)
4.6. We say that A has a finite presentation if there is a short exact sequence
0-> N -> P-> A'->O with P finitely-generated projective and N finitely-
generated. Show that
(i) if A has a finite presentation, then, for every exact sequence
O->R->S->A->O
with S finitely-generated, R is also finitely-generated;

(ii) if A has a finite presentation, it has a finite presentation with P free;
(iii) if A has a finite presentation every presentation O->N->P->A->O
with P projective, N finitely-generated is finite, and every presentation
O->N->P->A->O with P finitely-generated projective is finite;
(iv) if A has a presentation 0-> N l -> P l -> A ->0 with PI finitely-generated
projective, and a presentation 0-> N z-> Pz-> A ->0 with Pz projective, N z
finitely-generated, then A has a finite presentation (indeed, both the given
presentations are finite).
4.7. Let A = K(XI' ... , Xn , . •• ) be the polynomial ring in countably many in-
determinates Xl' ... , X n , ... over the field K. Show that the ideal I generated
by Xl' ... ,Xn , . ,. is not finitely generated. Hence we may have a presentation
O->N->P->A->O with P finitely generated projective and N not finitely-
generated.
5. Projective Modules over a Principal Ideal Domain
Here we shall prove a rather difficult theorem about principal ideal

domains. We remark that a very simple proof is available if one is content
to consider only finitely generated A-modules; then the theorem forms
a part of the fundamental classical theorem on the structure of finitely
generated modules over principal ideal domains.
Recall that a principal ideal domain A is. a commutative ring with-
out divisors of zero in which every ideal is principal, i.e. generated by
5. Projective Modules over a Principal Ideal Domain 27
one element. It follows that as a module every ideal in A is isomorphic

to A itself.
Theorem 5.1. Over a principal ideal domain A every submodule of
a free A-module is free.
Since projective modules are direct summands in free modules,
this implies
Corollary 5.2. Over a principal ideal domain, every projective module
is free.
Corollary 5.3. Over a principal ideal domain, every submodule of a
projective module is projective.
Proof of Theorem 5.1 . Let P = EB A j, where Aj = A, be a free module
jEJ
and let R be a submodule of P. We shall show that R has a basis. Assume J
well-ordered and define for every j E J modules
P(j) = EB Ai , Pu ) = EB Ai .
i<j i~j
Then every_element a E P(j)!lR may be written uniquely in the form (b, A)

where bEP(j) and AEA j. We defin~a homomorphism fj:P(j)!lR-+A
by fj(a) = A. Since the kernel of fj is PU)!lR we obtain an exact sequence
~j)!lR>-->PU)!lR-imfj .
Clearly imfj is an ideal in A. Since A is a principal ideal domain, this ideal

is generated by one element, say Aj . For Aj =1= 0 we choose cj E P(j)!lR,
such that fj(c) = Aj . Let J' ~ J consist of those j such that Aj =1= O. We
claim that the family {c j }, j E J', is a basis of R.
n
First we show that {cj},j E J', is linearly independent. Let L Il k h = 0

C.
k=1
and let jl <jz < ... <jn' Then applying the homomorphism fjn' we get
Iln fjJcjJ = IlnAjn = O. Since Ajn =1= 0 this implies Iln = O. The assertion then
follows by induction on n.
Finally, we show that {c j }, j E J' , generates R. Assume the contrary.
Then there is a least i E J such that there exists a E P (i)!lR which cannot
be written as a linear combination of {Cj},jEJ'. Ifi¢J', then aE ~i)!lR;
but then there exists k < i such that a E P(k)!lR, contradicting the mini-
mality of i. Thus i E J'.
Consider fi(a) = 11 Ai and form b = a - IlCi' Clearly
fi(b) = fJa) - fi(llcJ = O.
Hence b E ~i)!lR, and b cannot be written as a linear combination of
{c j }, j E J'. But there exists k < i with bE P(k)!lR, thus contradicting the
minimality of i. Hence {cj},j E J', is a basis of R. 0
28 I. Modules
Exercises:
5.1. Prove the following proposition, due to Kaplansky : Let A be a ring in which
every left ideal is projective. Then every submodule of a free A-module is
isomorphic to a direct sum of modules each of which is isomorphic to a left
ideal in A . Hence every submodule of a projective module is projective.
(Hint : Proceed as in the proof of Theorem 5.1.)
5.2. Prove that a submodule of a finitely-generated module over a principal ideal
domain is finitely-generated . State the fundamental theorem for finitely-
generated modules over principal ideal domains.
5.3. Let A, B, C be finitely generated modules over the principal ideal domain A.
Show that if A EB C ~ B EB C, then A ~ B. Give counterexamples if one drops
(a) the condition that the modules be finitely generated, (b) the condition that A
is a principal ideal domain.
5.4. Show that submodules of projective modules need not be projective. (A = 7l p 2,
where p is a prime. 7l p >--+71 p 2-71 p is short exact but does not split!)
5.5. Develop a theory of linear transformations T: V ---> V of finite-dimensional
vector spaces over a field K by utilizing the fundamental theorem in the
integral domain K[T].
6. Dualization, Injective Modules
We introduce here the process of dualization only as a heuristic

procedure. However, we shaH see in Chapter II that it is a special case of
a more general and canonical procedure. Suppose given a statement
involving only modules and homomorphisms of modules; for example,
the characterization ofthe direct sum of modules by its universal property
given in Proposition 3.2 :
"The system consisting of the direct sum S of modules {A j }, j E J,
together with the homomorphisms I j : Ar..... S' is characterized by
the following property. To any module M and homomorphisms
{1pj: Ar-..... M},jEJ, there is a unique homomorphism 1p: S---->M such that
for every j E J the diagram
is commutative."
The dual of such a statement is obtained by "reversing the arrows" ;
more precisely, whenever in the original statement a homomorphism
occurs we replace it by a homomorphism in the opposite direction.
In our example the dual statement reads therefore as foHows:
"Given a module T and homomorphisms {7rj: T---->A j }, jEJ. To
any module M and homomorphisms {qJj : M ----> A j}, j E J , there exists a
6. Dualization, Injective Modules 29
unique homomorphism cp: M - T such that for every j E J the diagram
T. ···················M
is commutative."
It is readily seen that this is the universal property characterizing
the direct product of modules {A j}, j E J, the nj being the canonical
projections (Proposition 3.3). We therefore say that the notion of the
direct product is dual to the notion of the direct sum.
Clearly to dualize a given statement we have to express it entirely
in terms of modules and homomorphisms (not elements etc.). This can be
done for a great many - though not all - of the basic notions introduced
in Sections 1, ... ,5. In the remainder of this section we shall deal with a
very important special case in greater detail: We define the class of
injective modules by a property dual to the defining property of projective
modules. Since in our original definition of projective modules the term
"surjective" occurs, we first have to find a characterization of surjective
homomorphisms in terms of modules and homomorphisms only. This
is achieved by the following definition and Proposition 6.1.
Definition. A module homomorphism c:: B- C is epimorphic or an
epimorphism if IXI c: = IX2 c: implies IXI = IX2 for any two homomorphisms
IXi:C-M, i=1,2.
Proposition 6.1. c:: B- C is epimorphic if and only if it is surjective.

Proof. Let B....'4c;;.' M. If c: is surjective then clearly IXI c:b = IX2 c:b
for all bE B, implies IXI c = IX2 C for all c E C. Conversely, suppose c: epi-
morphic and consider B....'4 C #
C/c: B, where n is the canonical projec-
tion and 0 is the zero map. Since Oc: = 0 = nc:, we obtain 0 = n and there-
fore Cjc:B=O or C=c:B. 0
Dualizing the above definition in the obvious way we have
Definition. The module homomorphism J1: A - B is monomorphic
or a monomorphism if J1IX I = J1IX2 implies IXI = IX2 for any two homo-
morphisms IX i : M-A, i = 1,2.
Of course one expects that "monomorphic" means the same thing
as "injective". For modules this is indeed the case; thus we have
Proposition 6.2. J1 : A - B is monomorphic if and only if it is injective.
Proof. If J1 is injective, then J1IX I x = J1IX2 x for all x E M implies
IXI x = IX2 X for all x E M. Conversely, suppose J1 monomorphic and
ai ' a2 E A such that J1a l = J1a 2 . Choose M = A and IX i : A-A such that
IXi(1) = ai' i = 1, 2. Then clearly J1IX I = J1IX2; hence IXI = IX2 and a l = a 2 · 0
30 I. Modules
It should be remarked here that from the categorical point of view

(Chapter II) definitions should whenever possible be worded in terms of
maps only. The basic notions therefore are "epimorphism" and "mono-
morphism", both of which are defined entirely in terms of maps. It is
a fortunate coincidence that, for modules, "monomorphic" and "injective"
on the one hand and "epimorphic" and "surjective" on the other hand
mean the same thing. We shall see in Chapter II that in other categories
monomorphisms do not have to be injective and epimorphisms do not
have to be surjective. Notice that, to test whether a homomorphism is
injective (surjective) one simply has to look at the homomorphism
itself, whereas to test whether a homomorphism is monomorphic
(epimorphic) one has, in principle, to consult all A-module homo-
morphisms.
We are now prepared to dualize the notion of a projective module.
Definition. A A-module I is called injective if for every homomorphism
!Y. : A ~ I and every monomorphism fJ.: A >-> B there exists a homo-
morphism {3 : B~ I such that {3 fJ. =!Y., i.e. such that the diagram
A~B
alp
1"/
is commutative. Since fJ. may be regarded as an embedding, it is natural
simply to say that I is injective if homomorphisms into I may be extended
(from a given domain A to a larger domain B).
Clearly, one will expect that propositions about projective modules
will dualize to propositions about injective modules. The reader must
be warned, however, that even if the statement of a proposition is dualiz-
able, the proof may not be. Thus it may happen that the dual of a true
proposition turns out to be false. One must therefore give a proof of the
dual proposition. One of the main objectives of Section 8 will, in fact,
be to formulate and prove the dual of Theorem 4.7 (see Theorem 8.4).
However, we shall need some preparation ; first we state the dual of
Proposition 4.5.
Proposition 6.3. A direct product of modules Il I j is injective if and
only if each Ij is injective. 0 j EJ
The reader may check that in this particular instance the proof of
Proposition 4.5 is dualizable. We therefore leave the details to the reader.
Exercises:
6.1. (a) Show that the zero module 0 is characterized by the property: To any
module M there exists precisely one homomorphism cp: O-+M.
(b) Show that the dual property also characterizes the zero module.
7. Injective Modules over a Principal Ideal Domain 31
6.2. Give a universal characterization of kernel and cokernel, and show that kernel
and cokernel are dual notions.
6.3. Dualize the assertions of Lemma 1.1, the Five Lemma (Exercise 1.2) and those
of Exercises 3.4 and 3.5.
6.4. Let rp : A -+ B. Characterize im rp, rp - \ Bo for Bo ~ B, without using elements.
What are their duals? Hence (or otherwise) characterize exactness.
6.5. What is the dual of the canonical homomorphism (J : EEl
Ai-+ IT
A;? What is
i eJ ie J
the dual of the assertion that (J is an injection? Is the dual true?
7. Injective Modules over a Principal Ideal Domain
Recall that by Corollary 5.2 every projective module over a principal

ideal domain is free. It is reasonable to expect that the injective modules
over a principal ideal domain also have a simple structure. We first
define:
Definition. Let A be an integral domain. A A-module D is divisible
if for every dE D and every 0 =1= A E A there exists C E D such that AC = d.
Note that we do not require the uniqueness of c.
We list a few examples:
(a) As Z-module the additive group of the rationals <Q is divisible.
In this example C is uniquely determined.
(b) As Z-module <Q/Z is divisible. Here c is not uniquely determined.
(c) The additive group of the reals IR, as well as IR/Z, are divisible.
(d) A non-trivial finitely generated abelian group A is never divisible.
Indeed, A is a direct sum of cyclic groups, which clearly are not divisible.
Theorem 7.1. Let A be a principal ideal domain. A A-module is in-
jective if and only if it is divisible.
Proof. First suppose D is injective. Let dE D and 0 =1= A EA. We
have to show that there exists c ED such that AC = d. Define rx: A---+D
by rx(l) = d and J1: A ---+ A by J1(1) = A. Since A is an integral domain,
J1(~) = ~ A = 0 if and only if ~ = O. Hence J1 is monomorphic. Since D is
injective, there exists f3: A ---+ D such that f3 J1 = rx. We obtain
d = rx(l) = f3 J1(1) = f3(A) = Af3(l).
Hence by setting C = f3(1) we obtain d = AC. (Notice that so far no use is

made of the fact that A is a principal ideal domain.)
Now suppose D is divisible. Consider the following diagram
A,.2......B
al
D
32 I. Modules
We have to show the existence of f3: B~ D such that f3 J1 = IY.. To

simplify the notation we consider J1 as an embedding of a submodule A
into B. We look at pairs (Aj,IY.j) with A~Aj~B, IY.j:Aj~D such that
IY.)A = IY.. Let rp be the set of all such pairs. Clearly rp is nonempty, since
(A, IY.) is in rp. The relation (A j ' IY.) ~ (Ak ' IY. k) if A j ~ Ak and IY.k IAj = IY. j
defines an ordering in rp. With this ordering rp is inductive. Indeed,
every chain (Aj' IY. j ), j E J has an upper bound, namely (U A j , UIY.)
where UAj is simply the union, and ~IY.{ is defined as follows: If a E UAj'
then a E Ak for some k E 1. We define U IY.j(a) = IY.k(a). Plainly UIY.j is well-
defined and is a homomorphism, and
(Aj' IY.) ~ (U A j , UIY.) .
By Zorn's Lemma there exists a maximal element (A, a) in rp. We shall
show that A = B, thus proving the theorem. Suppose A B; then there *
exists bE B with b tf A. The set of A E A such that Ab E A is readily seen
*
to be an ideal of A. Since A is a principal ideal domain, this ideal is genera ted
by one element, say AO. If Ao 0, then we use the fact that D is divisible
to find c E D such that a(Ao b) = Ao c. If AO = 0, we choose an arbitrary c.
The homomorphism a may now be extended to the module A generated
by A and b, by setting Ii (a + Ab) = a(a) + AC. We have to check that this
definition is consistent. If AbE A, we have Ii(Ab) = Ac. But A = ¢ AO for some
¢ E A and therefore Ab = ¢Aob. Hence
a(Ab) = a(¢ AO b) = ¢a(Ao b) = ¢ AOC = AC .
Since (A, a) < (A, Ii), this contradicts the maximality of (A, a), so that
A = B as desired. 0
Proposition 7.2. Every quotient of a divisible module is divisible.
Proof. Let e: D-- E be an epimorphism and let D be divisible.
For e E E and 0* A E A there exists dE D with e(d) = e and d' E D with
Ad' = d. Setting e' = e(d') we have Ae' = Ae(d') = dAd') = e(d) = e. D
As a corollary we obtain the dual of Corollary 5.3.
Corollary 7.3. Let A be a principal ideal domain. Every quotient of an
injective A-module is injective. 0
Next we restrict ourselves temporarily to abelian groups and prove
in that special case
Proposition 7.4. Every abelian group may be embedded in a divisible
(hence injective) abelian group.
The reader may compare this Proposition to Proposition 4.3, which
says that every A-module is a quotient of a free, hence projective, A-
module.
Proof. We shall define a monomorphism of the abelian group A
into a direct product of copies of (J).fll . By Proposition 6.3 this will
7. Injective Modules over a Principal Ideal Domain 33
*
suffice. Let 0 a E A and let (a) denote the subgroup of A generated by a.
*
Define a: (a)---+ (f).l'll as follows: If the order of a E A is infinite choose
0* a(a) arbitrary. If the order of a E A is finite, say n. choose 0 a(a)
to have order dividing n. Since (f).l'll is injective, there exists a map
f3a: A---+<Q/7l such that the diagram
(a )>-------+ A
a 1~ // Po
<Q{ll
is commutative. By the universal property of the product, the f3a define
a unique homomorphism f3 : A ---+ TI
(<Q{ll)a. Clearly f3 is a monomorphism
*
since f3a(a) 0 if a O. 0*
*
For abelian groups, the additive group of the integers'll is projective
and has the property that to any abelian group G 0 there exists a non-
*
zero homomorphism cp: 'll---+ G. The group <Q{ll has the dual properties;
it is injective and to any abelian group G 0 there is a nonzero homo-
morphism 1p: G---+<Q{ll. Since a direct sum of copies of'll is called free,
we shall term a direct product of copies of <Q/7l cofree. Note that the two
properties of'll mentioned above do not characterize'll entirely. Therefore
"cofree" is not the exact dual of "free", it is dual only in certain respects.
In Section 8 the generalization of this concept to ?rbitrary rings is
carried through.
Exercises:
7.1. Prove the following proposition : The A-module I is injective if and only if
for every left ideal J C A and for every A-module homomorphism IX : J --+ I the
diagram J>----> A
·L/p
I
may be completed by a homomorphism p : A --+ I such that the resulting triangle
is commutative. (Hint : Proceed as in the proof of Theorem 7.1.)
7.2. Let O--+R--+F--+A--+O be a short exact sequence of abelian groups, with F
free. By embedding F in a direct sum of copies of ~, show how to embed A
in a divisible group.
7.3. Show that every abelian group admits a unique maximal divisible subgroup.
7.4. Show that if A is a finite abelian group, then Homz(A, ~(Z) ~ A. Deduce
that if there is a short exact sequence O--+A'--+A--+A"--+O of abelian groups
with A finite, then there is a short exact sequence 0--+ A" --+ A --+ A' --+0.
7.5. Show that a torsion-free divisible group D is a ~-vector space. Show that
Homz(A, D) is then also divisible. Is this true for any divisible group D?
7.6. Show that ~ is a direct summand in a direct product of copies of ~(Z.
34 I. Modules
8. Cofree Modules
Let A be a right A-module and let G be an abelian group. Regarding A

as an abelian group we can form the abelian group Homz(A, G) of
homomorphisms from A into G. Using the right A-module structure
of A we define in Homz(A, G) a left A-module structure as follows:
(Acp)(a) = cp(aA), a E A, A E A, cp E Homz(A, G) .
We leave it to the reader to verify the axioms. Similarly if A is a left

A-module, Homz(A, G) acquires the structure of a right A-module.
Proposition 8.1. Let A be a left A-module and let G be an abelian group.
Regard Homz(A, G) as a left A-module via the right A-module structure
of A. Then there is an isomorphism of abelian groups
11 = 11A : HomA(A, Homz(A, G»)~ Homz(A, G).
Moreover, for every A-module homomorphism IX : A-B the diagram
HomA(B, Homz(A, G))~Homz(B, G)
1~. 1~. (8.1)
HomA(A, Homz(A, G»)~Homz(A, G)
is commutative. (In this situation we shall say that 11 is natural.)

Proof· Let cp : A-Homz(A, G) be a A-module homomorphism.
We define a homomorphism of abelian groups cp': A-G by
qj'(a) = (cp(a)) (1), a EA.
Conversely, a homomorphism of abelian groups 1p : A-G gives rise to
1p' :A-Homt(A, G) by (1p'(a»)(A)=1p(Aa), aEA, AEA. Clearly 1p' is a
homomorphism of abelian groups. We have to show that 1p' is a homo-
morphism of A-modules. Indeed, let (E A, then (1p'«(a»)(A) = 1p(A(a);
on the other hand (((1p'(a»)) (A) = (1p'(a») (A() = 1p(A(a). Clearly, cp'r-+cp'
and 1p'r-+1p' are homomorphisms of abelian groups. Finally, we claim
(cp')'=cp and (1p')'=1p. Indeed, (1p')'(a)=(1p'(a»)(1)=1p(a), and
(cp')'(a») (A) = cp'(Aa) = (cp(Aa») (1),
but
(cp(Aa») (1) = (A (cp(a»)) (1) = (cp(a») (1 A) = (cp(a») (A) ,
since cp is a A-module homomorphism. Thus we define 11 by setting

11(CP) = cp', and 11 is an isomorphism. The naturality of 11, i.e. the com-
mutativity of the diagram (8.1), is evident. Notice that IX* on the right
of the diagram (8.1) is a homomorphism of right A-modules. 0
8. Cofree Modules 35
We now look at A* = Hom 71 (A, <Q/Z), which is made into a left

A-module using the right A-module structure of A. We claim that A*
has the property that to any nonzero A-module A there is a nonzero
homomorphism <p: A - A *. Indeed, any nonzero homomorphism of
abelian groups 1p: A- <Q/ll will correspond by Proposition 8.1 to a
nonzero <p: A-A*. Also, it will follow from Theorem 8.2 below that
A* is injective (set G = <Q/Z). We therefore define
Definition. A A-module is cofree if it is the direct product of modules
A* = Hom71(A, <Q/Z). Note that this is consistent with the description
of <Q/Z as a co free group, since Hom 71 (Z, <Q/Z) ~ <Q/Z.
Theorem 8.2. Let G be a divisible abelian group. Then A = Hom 71 (A, G)
is an injective A-module.
Proof. Let J.i: A-B be a monomorphism of A-modules, and let
ex: A-A a homomorphism of A-modules. We have to show that there
exists 13 : B- A such that 13 J.i = ex. To prove this, we remark that ex : A - A
corresponds by Proposition 8.1 to a homomorphism of abelian groups
ex': A-G. Since G is injective, there exists 13': B-G such that 13' J.i = ex'.
Under the inverse of 1'/ in Proposition 8.1 we obtain a homomorphism of
A-modules 13: B- X. Finally by the naturality of 1'/, the diagram
is commutative. 0
We are now prepared to prove the dual of Proposition 4.3.
Proposition 8.3. Every A-module A is a sub module of a cofree, hence
injective, A-module.
*
Proof. Let 0 a E A and let (a) denote the submodule of A generated
by a. By the remarks preceeding Theorem 8.2 there exists a nonzero
A-homomorphism ex : (a)-A *. Since A * is injective there exists f3a: A -A*
such that the diagram
is commutative. By the universal property ofthe direct product the f3a define
a homomorphism 13: A - n(A:), where A: = A *. Clearly 13 is mono-
morphic. 0
I. Modules
36
We conclude this section by dualizing Theorem 4.7.

Theorem 8.4. For a A -module I the following statements are equivalent:
(1) I is injective;
(2) for every exact sequence AhB~ C of A-modules the induced
sequence
is exact;
(3) if Ji : 1>--> B is a monomorphism, then there exists f3 : B-- I such that
f3Ji= 1/;
(4) I is a direct summand in every module which contains I as sub-
module;
(5) I is a direct summand in a cofree module.
The proof is dual to the proof of Theorem 4.7. For the step (3) => (4)
one needs the dual of Lemma 4.6. The details are left to the reader. 0
Note that, to preserve duality, one should really speak of "direct
factor" in (4) and (5), rather than "direct summand". However, the two
notions coincide!
Exercises:
8.1. Complete the proof of Theorem 8.4.
8.2. Let A be a A-module and let G be a divisible abelian group containing A.
Show that we may embed A in an injective module by the scheme
A = HomA(A, A) ~ Homz(A, A) ~ Homz(A, G).
(You should check that we obtain an embedding of A-modules.)
8.3. For any A-module A, let A* be the right A-module Homz(A, <Q/Z). Show
that A is naturally embedded in A**. Use this embedding and a free presenta-
tion of A * to embed A in a cofree module.
8.4. Suppose given 0-> A-->ll->JI -> 0, 0->A->I2 ->J2 ->0, with 11 ,12 injective.
Show that II (f;J2 ~ 12 (f;JI • To what statement is this dual?
8.5. State a property of injective modules which you suspect may not hold for
arbitrary divisible modules.
9. Essential Extensions
In this section we shall show that to a given A-module A there exists

an injective module E containing A such that every injective module
containing A also contains an isomorphic copy of E. This property
will define E up to isomorphism. E is called the injective envelope of A.
We remark (see Exercise 9.5) that the dual of the injective envelope
("projective cover") does not exist in general.
Definition. A monomorphism Ji : A >--> B is called essential if for any
submodule H of B, H =1= 0 implies H nJiA =1= O. If A is regarded as a sub-
module of B then B is called an essential extension of A (see [12J).
9. Essential Extensions 37
Examples. (a) As abelian group <Q is an essential extension of 71..

(b) The module B = A EB C can never be an essential extension of A,
unless C=O. For CnA=O.
Note that if B is an essential extension of A, and C is an essential
extension of B, then C is an essential extension of A.
Proposition 9.1. B is an essential extension of A if and only if, for every
0* bE B , there exists A E A such that Ab E A and Ab * O.
* *
Proof. Let B be an essential extension of A, and let H be the submodule
generated by bE B. Since H 0 it follows that H nA 0, i.e. there
exists A E A such that 0 =+= Ab E A. Conversely, let H be a nontrivial
* *
submodule of B. For 0 hE H there exists A E A such that 0 Ah E A.
Therefore H nA 0, and B is an essential extension of A. 0
*
Let A be a submodule of a A-module M. Consider the set of essential
extensions of A, contained in M. Since A is an essential extension of itself,
 is not empty. Under inclusion, is inductive. Indeed, if {E j }, j EJ,
is a chain of essential extensions of A contained in M , then it follows
easily from Proposition 9.1 that their union U E j is again an essential
jeJ
extension of A contained in M. By Zorn's Lemma there exists a maximal
essential extension E of A which is contained in M .
Theorem 9.2. Let A be a submodule of the injective module I. Let E

be a maximal essential extension of A contained in I. Then E is injective.
Proof. First we show that E does not admit any nontrivial essential
monomorphism.
Let /1: E-X be an essential monomorphism. Since I is injective,
there exists a homomorphism ¢ : X - I completing the diagram
E~X
I //
I '"
We show that ¢ is monomorphic. Let H be the kernel of ¢. We then

have H r;. X and H n /1E = O. Hence ker ¢ = H = 0, for /1 is essential. It
follows that ¢X is an essential extension of A contained in I. Since E is
maximal, it follows that X = E.
Now consider the set tp of submodules Hr;.1 such that HnE=O.
Since 0 E 'P, 'P is nonempty, it is ordered by inclusion and ind~ctive.
Hence by Zorn's Lemma there exists a maximal submo~ule H of I
such that Hn E = O. The canonical projection n: I -- I IH induces a
monomorphism a=nIE:E>--+II H. We shall show that a is essential.
Let HIB be a nontrivial submodule of l iB, i.e. let Ii c H ~ I where the
I. Modules
38
first inclusion is strict. By the maximality of H the intersection H n E

is nontrivial, hence H/B n (jE is nontrivial. It follows that (j is essential.
By the first part ofthe proof E admits no proper essential monomorphism,
whence it follows that (j: E~I/H is an isomorphism. The sequence
H >---> I ~ E now splits by the embedding of E in I. Therefore E is
a direct summand in I and is injective by Proposition 6.3. 0
Corollary 9.3. Let E 1 , E2 be two maximal essential extensions of A
contained in injective modules II' 12, Then El ~ E2 and every injective
module I containing A also contains a submodule isomorphic to El .
Definition. El is called the injective envelope of A.
Proof. Consider the diagram
r~E'
E2
Since E2 is injective there exists ~: El ~ E2 completing the diagram.
As in the proof of Theorem 9.2 one shows that ~ is monomorphic. But
then E 2 , as an essential extension of A, is also an essential extension of E 1 ,
which shows, again as in the proof of Theorem 9.2, that ~: El ~ E 2·
The proof of the second part is now trivial. 0
Exercises:
9.1. Compute the injective envelope of?l, ?lp' p prime, ?In.
9.2. Show that if Bi is an essential extension of Ai' i = 1,2, then BI EB Bl is an
essential extension of Al EBA 1 . Extend this to direct sums over any index set J.
9.3. Given any abelian group A, let T(A) be its torsion subgroup and F(A) = A / T(A).
Show that cp : A ->B induces T(cp): T(A)-> T(B), F(cp): F(A)->F(B), and that cp
is a monomorphism if and only if T(cp) and F(cp) are monomorphisms. Show
that the monomorphism cp is essential, if and only if T(cp) and F(cp) are essential.
Now suppose given
0-----> T( A) -----> A -----> F (A) -----> 0
I I I
0--> T(B)--> B --> F (B)--> 0
I I I
0--> T(C)--> C--> F(C)--> 0
where A ~ C is to be regarded as fixed, and B is an essential extension of A.

Show that if T(B) and F(B) are maximal, so is B. Show that if B is maximal,
so is T(B), but that F(B) may fail to be maximal.
9. Essential Extensions 39
Show that if C is divisible, so are 7{C) and F(C). What does this tell us
about the injective envelope of T(A), A and F(A)?
9.4. Give a procedure for calculating the maximal essential extension of A in B,
where B is a finitely generated abelian group.
9.5. Show that the dual of an injective envelope does not always exist. That is, given
a A-module A , we cannot in general find P-4.A ,P projective, such that, given
Q~A, Q projective, we may factor '1 as Q~P"":""A. (Hint : Take A =71,
A = 715.) Where does the dual argument fail?

Module Collected

Uploaded by

Copyright:

Available Formats

You might also like

Module Collected

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Module Collected

Uploaded by

Copyright:

Available Formats

Chapter 3

Modules and Vector Spaces

3.1 Definitions and Examples

(1.1) Definition. Let R be an arbitrary ring with identity (not necessarily

We leave it as an exercise to check that : R(G) - R(G) is an

(1.3) Definition. Let R be a ring and let M, N be R-module8. A function

The set of all R-module homomorphisms from M to N will be de-

f (ng) = f (9 + ... + 9) = f (9) + ... + f(9) = nf(9)

Mm(R) x M,,,,"(R) Mm."(R)

(1.1) r(mlm2) = (rmi)m2 = ml(rm2).

For example, every ring is a Z-algebra, and if R is a commutative ring,

M are R-modules then HomR(M, N) is an abelian group via

(af)(rm) = a(f (rm)) = a(r(f (m))) = ar f (m).

This last expression is equal to r(a f) (m) = ra f (m) if R is a commu-

f (X) =ao+a1X +...+aX"

f(X) . (U1, U2) = f(T)(u1,u2)

3.2 Submodules and Quotient Modules

(2.1) Lemma. If M is an R-module and N is a nonempty subset of M,

If F is a field and V is a vector space over F, then an F-submodule of

(2.3) Lemma. Let M be an R-module and let (NQ}QEA be a family of sub-

(2.4) Theorem. (First isomorphism theorem) Let M and N be modules

Proof. Let K = Ker(f). From Theorem 1.3.10 we know that f : M/K -

(2.5) Theorem. (Second isomorphism theorem) Let M be an R-module and

(2.6) Theorem. (Third isomorphism theorem) Let M be an R-module and

Proof. Define f : Al/P -. MIN by f (m+ P) = m+N. This is a well-defined

Ker(f)={m +P:m+N=N}={m +P:mEN}=NIP.

(2.7) Theorem. (Correspondence theorem) Let Al be an R-module, N a

Thus, (S) is a submodule of M that contains S and it is contained in

(2.9) Lemma. Let M be an R-module and let S C M. If S = 0 then (S) =

(2.10) Definition. We say that the R-module M is finitely generated if

If M is a finitely generated R-module and N is any submodule, then

(2.12) Proposition. Suppose M is an R-module and N is a submodule. If

alyl+ +atyt E (T). Then ir(x-y) =O so that x-y E Ker(ir) = N = (S).

(2.14) Definition. If R is a ring, M is an R-module, and X is a subset of

It is easy to check that Ann(X) is a left ideal of R, and furthermore,

Example. Let F be a field, V a vector space, T E EndF(V) a linear trans-

(2.15) Proposition. Let R be a ring and let M = (m) be a cyclic R-module.

(2.16) Corollary. If F is a field and M is a nonzero cyclic F-module then

If M is an R-module and I C R is an ideal then we can define the

The set IM is easily checked to be a submodule of M. The product IM

(2.17) Definition. Let R be an integral domain and let M be an R-module.

(2.18) Proposition. Let R be an integral domain and let M be an R-module.

Proof. (1) Let x, y E M, and let c, d E R. There are a 34 0, b 36 0 E R such

(2.20) Proposition. Let R be an integral domain and let M be a finitely gen-

3.3 Direct Sums, Exact Sequences, and Hom

The direct sum has an important homomorphism property, which, in

f is an R-module homomorphism and it follows from condition (1) that f is

a family of R-modules indexed by the (possibly infinite) set J. Then the

(xj)jEJ'+' (yj)JEJ = (x, +Yj)jE

®Mj = {(xj)jEJ : xj = 0 except for finitely many indices j E J}.

It is easy to check that ®jEJMj is a submodule of IIjEJ Mj.

(3.2) Theorem. Let M be an R-module and let {Mj}jEJ be a family of

(3.3) Definition. If M is an R-module and MI C M is a submodule, we say

(3.4) Example. Let R = Z and M = Z e. If MI = (p) then MI is not

The concept of exact sequences of R-modules and R-module homo-

(3.5) Definition. Let R be a ring. A sequence of R-modules and R-module

is said to be exact at Mi if Im(fi) = Ker(fi+I ). The sequence is said to be

As particular cases of this definition note that

(3.1) 0 - . MI _ M 9-+ M2'0