§ 4. Finitely generated modules over a principal ideal domain
1. Direct sums of cyclic modules
Let A be a commutative ring. Recall (II, p. 220, Prop. 22) that a cyclic A-module is isomorphic to a quotient module , where a is an ideal of A. We will see later (Sect. 4) that every finitely generated module over a principal ideal domain is a direct sum of finitely many cyclic modules.
Proposition 1
— *Let E be a module over a commutative ring A; suppose E is a direct sum of n cyclic modules (), where the are ideals of A;
then, for each integer , the -module is isomorphic to the direct sum of the modules , where for each -element subset of , the ideal is .
Let be the canonical generator of , that is the image of the unit element of , so that is the direct sum of the (). Then we know (III, p. 515, Prop. 10) that the exterior algebra is isomorphic as an -module to the tensor product . Now is just the direct sum , since the exterior product of any two elements of is zero, and so is the direct sum of the modules as runs over the set of -element subsets of (with ) ; now is known to be isomorphic to (II, p. 257, Cor. 4), which completes the proof.
We now see that, in the notation of Prop. 1, if the ideals form an increasing sequence, they are completely determined by the module . More precisely :
Proposition 2
— *Let be a commutative ring, and let be a direct sum of cyclic modules , where the satisfy . Then, for , the ideal is the annihilator of ; if then for and for .
Indeed, in the notation of Prop. 1, we have , where denotes the greatest element of the subset . Since for every -element subset , and since for , it follows that is the intersection of , as varies over the set of -element subsets of ; the ideal is thus indeed the annihilator of , by Prop. 1.
Corollary
— In the notation of Prop. 2, if , and if is also isomorphic to the direct sum of cyclic modules with , then and for (uniqueness of the ).
2. Content of an element of a free module
Let be a principal ideal domain, let be a free -module, and let be an element of . As runs through the set of linear forms on , the elements form an ideal of , called the content of in . An element of is called a content of in if it generates the ideal ; this amounts to saying that there exists a linear form on such that and that divides for every linear form on . Let be a basis of ; put ; then the ideal consists of sums , as runs through the set ; it follows immediately that an element of is a content of in if and only if it is a gcd of the family of coordinates of .
We say that is indivisible if , that is if the coordinates of with respect to a basis of are setwise coprime.
Lemma 1
Let be a free module over a principal ideal domain and let be an element of . Then the following conditions are equivalent*:
(i) is indivisible ; (ii) *there exists a linear form on such that ; (iii) is nonzero and the submodule is a direct factor of ; (iv) is an element of some basis of .
(i) (ii) : this follows from the definition. (ii) (iii) : let be a linear form on such that ; then and the map is a projector of , with image . (iii) (iv) : let be a complement of in , and let be a basis of (VII, p. 15, Cor. 2) ; then is a basis of . (iv) (i) : trivial.
Remark
— 1) If is a nonzero element of and a content of , then there exists a unique element of such that ; denote this element ; then is an indivisible element of . 2) The content is the annihilator of the torsion module of .
Let be a free module over a principal ideal domain and let be a submodule of ; by VII, p. 4, Lemma 1, the family , has a maximal element ; if then such a maximal element is nonzero.
Proposition 3
Let be a free module over a principal ideal domain and let be a nonzero submodule of . Let be an element of such that is maximal among the contents of elements of , let be a content of in , and let be a linear form on such that .
a) is the direct sum of and the kernel of . b) is the direct sum of and . c) for every linear form on .
Put ; clearly , since . Furthermore, for each we have
b) we have . To prove c), it is therefore sufficient to prove that for every linear form on , or equivalently for every linear form on such that ; now, if and is a gcd of and , then there exist with ; then , which implies as above that , and c) follows.
3. Invariant factors of a submodule
Theorem 1
— Let be a free module over a principal ideal domain , and let be a submodule of of finite rank . Then there exist a basis of , elements of , and nonzero elements of () such that: a) the form a basis of ; b) divides , for .
Moreover the module generated by the and the principal ideals are uniquely determined by the above conditions; the module is the torsion submodule of , and is isomorphic to the direct sum of the -modules ; finally is the direct sum of and a free module isomorphic to .
1) Existence of the and the .
If then the theorem is trivial. If then it follows from Prop. 3 that there exist an element of , a nonzero element , of , and a submodule of , such that is the direct sum of and , such that is the direct sum of and the submodule of , and such that for every linear form on .
Now we can proceed by induction on the rank of . Since is a free module (VII, p. 15, Cor. 2) and has rank , there exist a basis of , elements of , and nonzero elements , of such that is a basis of , and divides for . If is the submodule of , generated by the elements of distinct from , then is the direct sum of and the module generated by ; now is a basis for , and is a basis for . It now remains only to show that divides ; but has the form , where is the linear form on defined by for , and ; and we have seen above that .
2) Uniqueness properties.
Since the are nonzero, the module is the set of such that for some in ; in other words is the torsion submodule of . This uniquely determines .
It is clear that is isomorphic to the direct sum of the cyclic modules (II, p. 204, formula (26)). Let be the number of the ideals , which are distinct from : so that the first ideals are equal to , and the last rare distinct from . Then is also isomorphic to the direct sum of the modules , where .
Thus the conditions of the Cor. to Prop. 2 (VII, p. 16) are satisfied: the ideals () are thus uniquely determined by the module .
Since is the direct sum of and , it follows that is the sum of and , and this sum is direct since ; on the other hand is isomorphic to (I, p. 41, Th. 4, c)), that is to , which shows that is a free module isomorphic to .
Corollary
— A submodule of finite rank in a free module over a principal ideal domain has a complement in if and only if is torsion free.
In the notation of Th. 1, if is torsion free, then , and has a complement in . Conversely if has a complement in , then is isomorphic to , which is free (VII, p. 15, Cor. 2), and a fortiori torsion free.
Remark
It can happen that a submodule of infinite rank in a free module can be such that is torsion free, but has no complement in (VII, p. 60, Ex. 6, b)).
Definition 1
— In the notation and hypotheses of Th. 1, the ideals of are called the invariant factors of the submodule with respect to the module .
In the case where is either the ring of rational integers or the polynomial ring in one indeterminate over a field , there is a canonical way of choosing a generator for each ideal of : a positive integer in the case of , or a monic polynomial in the case of (VII, p. 5). In each of these cases, the canonical generator of the invariant factor is also called an invariant factor of with respect to , by abuse of language.
4. Structure of finitely generated modules
Theorem 2
— Every finitely generated module over a principal ideal domain is isomorphic to a direct sum of a finite number of cyclic modules , where the are ideals of (some of which may be zero) such that , and which are uniquely determined by these conditions.
If can be generated by elements then it is isomorphic to a quotient module , where (II. p. 218). Since has finite rank (VII, p. 15, Prop. 1), the conditions of Th. 1 (VII, p. 18) are satisfied. Then in the notation of Th. 1 is isomorphic to a direct sum of a complement of in and the torsion module . The module is free of finite rank , so isomorphic to . If is the smallest index such that , then is isomorphic to the direct sum of modules for . The stated conditions will then be satisfied if we take , for , and for . Uniqueness follows from the Cor. of VII, p. 16.
Corollary 1
— *Every finitely generated module E over a principal ideal domain is the direct sum of the torsion submodule of E and a free module.
The torsion submodule of E has in general several distinct complements. For example, if then the torsion submodule of E is ; it has as a complement the submodule , and also the submodule consisting of all elements , where runs through and is the residue class of mod 2.
Corollary 2
— Every torsion free finitely generated module over a principal ideal domain is free of finite rank. This follows immediately from Cor. 1.
The condition that the module be finitely generated is essential. For example the additive group of the field of fractions K of A is torsion free as an A-module; however it is not a free module if , for on the one hand any two elements of K have a common divisor, and on the other hand K is not a cyclic A-module, for otherwise (), whence (), so , and .
Definition 2
— In the notation and hypotheses of Th. 2, the ideals are called the invariant factors of the module E. As in Def. 1 (VII, p. 19), when or , the canonical generator of the ideal (a positive integer or a monic polynomial) is also called, by abuse of language, an invariant factor of the finitely generated module E.
We must be careful not to confuse the invariant factors of a module E with those of a submodule M of a free module L with respect to the module L (Def. I).
5. Calculation of invariant factors
Proposition 4
— *Let A be a principal ideal domain, let L be a free A-module with finite basis (), let M be a submodule of L, let be a system of generators of M and let () be the invariant factors of M with respect to L. Then for the product is a gcd of the m-th order minors of the matrix whose columns are the coordinate vectors of the with respect to the basis .
By Th. 1 it is clear that ; hence the coordinates of any element of M are multiples of . On the other hand, there exists an element of M for which is a content in L. Expressing as a linear combination of the , it follows that is an element of the ideal generated by the coordinates of the . As these are all multiples of , it follows that is indeed their gcd, and our assertion is proved in the case .
For general , consider the m-th exterior power of M (III, p. 507). In the notation of Th. 1 there exists a basis for M where (); thus has a basis consisting of the elements , as runs through the set of strictly increasing sequences of elements of . Now the elements belong to a basis of . Hence the canonical map from into is an isomorphism from onto the submodule of having as a basis the elements , and we identify this submodule with . Since is a multiple of for , the elements are all multiples of , and one of them is equal to ; thus is a gcd of the coordinates with respect to of the elements of a system of generators of . The first part of the argument shows that then is a gcd of the set of coordinates of any system of generators of , with respect to any basis of . Taking the basis for induced from the basis of , and the system of generators for consisting of exterior products of the , the expression for the coordinates of these products in terms of determinants (III, p. 528, Prop. 9) gives the stated result.
6. Linear mappings of free modules, and matrices over a principal ideal domain
Let be a principal ideal domain. Consider a linear mapping from a free -module of rank into a free -module of rank . The preceding results allow us, by choosing suitable bases for and , to put the matrix for in a particularly simple form, called the canonical form of the matrix.
Proposition 5
Let be a principal ideal domain, and a linear map of rank from a free -module of rank into a free -module of rank . Then there exist bases () of and () of such that for and for , where the are nonzero elements of , each of which divides the next; the ideals are the invariant factors of in , and are thus uniquely determined.
Let be the kernel of ; the quotient is isomorphic to the module , which is a submodule of and so free (VII, p. 15, Cor. 2); thus has a complement in (II, p. 218, Prop. 21), and the restriction of to is an isomorphism from onto . If the ideals () are the invariant factors of in , then Th. 1 of VII, p. 18 shows that there exists a basis () of such that () is a basis of . Since the restriction of to is an isomorphism from onto , there exists a basis () of such that . This basis extends to a basis () of by taking () to be a basis of the kernel .
Corollary 1
— Let X be a matrix of rank r, with n rows and m columns, over a principal ideal domain A; then there exists a matrix equivalent to X (II, p. 354) of the form
where the are nonzero elements of A, each of which divides the next. Under these conditions the are uniquely determined up to multiplication by invertible elements.
Given that two matrices X and are equivalent if there exist invertible square matrices P and Q, of orders n and m, over A, such that , Cor. 1 is just Prop. 5 expressed in terms of matrices.
In the notation of Prop. 5 and Cor. 1, the (nonzero) ideals , are called the invariant factors of the linear map f, or of the matrix X. It then follows immediately from Cor. I that:
Corollary 2
— Two matrices X and with n rows and m columns over a principal ideal domain A are equivalent if and only if they have the same invariant factors.
Note that when A is a field we can take the to be equal to 1, and then we recover Prop. 13 of II, p. 360.
If X is the matrix of the linear mapping f with respect to an arbitrary basis of L and an arbitrary basis of L', then the columns of X are the coordinate vectors, with respect to the basis of L', of elements of L' which form a system of generators for . The following result is thus an immediate consequence of Prop. 4.
Proposition 6
— Let X be a matrix of rank r over a principal ideal domain A, and let () be its sequence of invariant factors. Then is a gcd of the elements of X; and the product is a gcd of the q-th order minors of X for .
7. Finitely generated abelian groups
In the case , the results of section 4 can be expressed
Theorem 3
— Every finitely generated abelian group G is a direct sum of its torsion subgroup F (subgroup of elements of finite order in G) und a free abelian group of finite rank p (isomorphic to ). The group F is a direct sum of a finite number of cyclic groups of orders , where the are integers , each of which divides the previous one; moreover, the integers and () are uniquely determined by .
Remark
While the orders of the cyclic groups of which F is the direct sum are well defined by the divisibility condition of Th. 3, it is not the same for the groups themselves: for example, in the product G of with itself (for p prime), the subgroups are precisely the -vector subspaces, and G is the direct sum of two 1-dimensional subspaces in different ways.
Corollary 1
— *In a finite abelian group G, there exists an element whose order is the lcm of all the orders of elements of G; this order n is the first invariant factor of G.
Corollary 2
— Any finite abelian group G whose order is not divisible by the square of any integer is cyclic. Let us keep to the notation of Th. 3. Then since G is finite, and , for otherwise the order of G would be divisible by . Hence G is cyclic.
Corollary 3
— *Let L and M be two free -modules of rank n, let be a basis for L and a basis for M (), let u be a homomorphism from L into M, and let U be its matrix with respect to the bases and . Then is finite if and only if , and then . By changing bases in L and M if necessary, we may assume that U has the form described in VII, p. 22, Cor. 1 to Prop. 5 (where the are in this case integers); the corollary then becomes obvious, since the order of a direct sum of -modules () is infinite if one of the is zero, and is equal to otherwise.
8. Indecomposable modules. Elementary divisors
Definition 3
— A left module M over a ring A is said to be decomposable if it is the direct sum of a family of proper nonzero submodules. Otherwise it is said to be indecomposable. The zero module is thus decomposable, being the direct sum of the empty family of submodules. Let be a left ideal of the ring A; the submodules of are just the quotients , where is an ideal of A containing (I, p. 41, Th. 4); if and are two ideals of A containing , then the module is the direct sum of the submodules and if and only if and . As a result:
Lemma 2
— *The module is indecomposable if and only if and there is no pair of ideals of A, distinct from A and , such that and .
Proposition 7
— Let be a commutative ring, let be a prime ideal of (I, p. 117, Def. 3), and let be an ideal of contained in . Suppose that for every there exists an integer such that . Then the -module is indecomposable.
Let and be two ideals of , such that and . Then ; if and , then , so (I, p. 316, Prop. 4); hence either or . Suppose for example that , so ; then there exist and such that ; let be such that ; then . so , so . Lemma 2 now shows that is indecomposable.
Now suppose that is a principal ideal domain; by VII, p. 2, Prop. 2, the prime ideals of are the ideals , where is an irreducible element of , and the ideal ; by the previous proposition, the modules and , for irreducible and , are indecomposable. Since every cyclic module is a direct sum of modules of this type (VII, p. 3, Prop. 4) and since every finitely generated -module is a direct sum of cyclic modules (VII, p. 19, Th. 2), we deduce:
Proposition 8
— Let be a principal ideal domain and let be a finitely generated -module.
a) is indecomposable if and only if it is isomorphic to or to a module of the form , where is an irreducible element of and is an integer.
b) is a direct sum of a finite family of indecomposable submodules.
Part b) of the above proposition can be made more precise as follows:
Proposition 9
— Let be a principal ideal domain, let be a system of representatives of irreducible elements of and let be a finitely generated -module. Then there exist positive integers and (), uniquely determined by and zero except for finitely many of them, such that is isomorphic to the direct sum of and the ().
The existence of the integers and () follows from Prop. 8. The integer is uniquely determined: it is the rank of the free module which is the quotient of by its torsion submodule. Finally, the -primary component of is isomorphic to the direct sum of the ; since the family of ideals () is totally ordered by inclusion, the uniqueness of the follows from the Cor. to Prop. 2 of VII, p. 16.
Definition 4
— In the notation of Prop. 9, those ideals ( an integer) such that are called the elementary divisors of the module , and the integers are called their multiplicities; if the integer is , it is called the multiplicity of the elementary divisor 0.
As for the invariant factors (VII, p. 19, Def. 1), when or (K a commutative field), then the canonical generator of the ideal (a positive integer or a monic polynomial) is also called, by abuse of language, an elementary divisor of the finitely generated module .
Remark 1
If is a finite abelian group, then its structure can be described by writing down its elementary divisors, each repeated as often as its multiplicity. We will say, for example, that is « of type » (or that it is « a group ») if it is isomorphic to the product of two groups , one group , two groups and one group .
Remark 2
If a finitely generated torsion module over a principal ideal domain is given as a direct sum of cyclic modules isomorphic to (in particular if the invariant factors of are known), then the elementary divisors of , and their multiplicities, can be determined by noticing that is isomorphic to the product of the , where is the decomposition of into irreducible factors (, p. 3). Let us study for example the multiplicative group , where denotes the multiplicative group (, p. 12). Since , this group is isomorphic to the product of the groups , , and (, p. 13, Th. 3); now the last three groups are cyclic of orders 20, 22 and 100, and is the product of two cyclic groups of order 2 (\emph{loc. cit.}); since , and , the group is of type .
Remark 3
To calculate the invariant factors of a torsion module whose elementary divisors are known, we again lean on the fact that, if the are pairwise coprime elements of , then the product is a cyclic module isomorphic to (, p. 3, Prop. 4). Let us illustrate the method by looking at the example of the group : write the elementary divisors of which are powers of the same irreducible on the same line, beginning with those of greatest exponent; extend these lines to lines of equal length by putting in 1's where necessary:
Then the invariant factors are the products of elements in the same column: . Indeed is isomorphic to a product of cyclic groups of orders 1100, 20, 2, 2, 2 by Prop. 4 of , p. 3; since each of these orders is a multiple of the next, these are the invariant factors of (, p. 22, Th. 3).
Remark 4
An -module is called simple (, p. 37) if it is nonzero and has no submodules other than itself and ; it is then necessarily cyclic, so finitely generated, and indecomposable; since the modules are not simple for , while the modules are, and since is simple only if the ring is a field, we deduce that the simple -modules are: a) free modules of rank 1, when is a field; b) modules isomorphic to quotients , where is an irreducible element of , when is not a field.
9. Duality in modules of finite length over a principal ideal domain
In this section denotes a principal ideal domain which is not a field (and hence has at least one irreducible element), and the field of fractions of . For every -module , put
we know that is equipped with the structure of an A-module in a natural way, namely for every homomorphism and every , the homomorphism maps to . To every homomorphism of A-modules is associated the homomorphism , where (II, p. 196). For and , we put ; then is an A-bilinear map from into , called canonical.
If M and are two A-modules, then to every A-bilinear map are associated the A-linear maps and , where (II, p. 268, Cor. to Prop. 1). In particular the canonical A-bilinear map defines an A-linear map (also called canonical)
such that for and .
Proposition 10
— If M is an A-module of finite length, then is (in general non-naturally) isomorphic to M, and the canonical map is an isomorphism. Using VII, p. 19, Th. 2 and II, p. 203, Cor. 1, we reduce to the case where is cyclic. Thus we may suppose that with . Note that any homomorphism is completely determined by the image under of the class of 1 mod , and this element must satisfy the relation ; conversely, for any such that , there exists a homomorphism such that . It follows that is isomorphic to , and since the homothety by is a bijection on K, we also have isomorphic to , which proves the first assertion. This proves that and are isomorphic, so have the same length ; on the other hand is injective, for if is such that the relation (for ) implies , then taking we have . It follows that the image is necessarily equal to .
Corollary
— Let M, N be two A-modules of finite length, and let be an A-bilinear map from into , such that : 1) the relation for all implies ; and 2) the relation for all implies . Then the A-linear maps and associated to are isomorphisms. Indeed the hypotheses on mean that and are injective and since and by Prop. 10, this implies that , and consequently and are bijective.
Proposition 11
— If is an exact sequence of A-modules of finite length, then the sequence is exact [^1].
Let us show first of all that given an exact sequence
(1)
the corresponding sequence
is exact; indeed we know that the sequence
is exact (II, p. 227, Th. 1); on the other hand, it follows from (1) that
(II, p. 212, Prop. 16); by Prop. 10 we thus have
in other words . Since is naturally identified with a submodule of , it must be equal to , which proves our assertion.
This implies immediately that if is injective, then is surjective; the conclusion then follows from II, p. 199, remark 4.
For every A-module M, let denote the set of submodules of M. For every submodule N of M (resp. every submodule N' of D(M)), let (resp. ) denote the submodule of (resp. M) consisting of those (resp. ) such that for all (resp. for all ).
Proposition 12
Let M be an A-module of finite length. Then the map which sends every submodule N of M to is a bijection from onto , and the inverse bijection sends every submodule N' of D(M) to the submodule of M; the module is naturally identified with and with . Moreover, we have
(2)
for all submodules of M.
For each submodule N of M, there is an exact sequence
and hence (Prop. 11) an exact sequence
and since the image of in is obviously , we see (Prop. 10) that ; since is identified with by Prop. 10, we have similarly
in addition it is clear that , so . Furthermore the first relation in (2) is obvious, and by applying it to the submodules and of , we have , so . This completes the proof of the proposition.
Example 1
For , the -modules of finite length are precisely the finite abelian groups; then , so . Then to define , we sometimes take, instead of , a -module isomorphic to it, such as (V, p. 79, Prop. 2) the group of roots of unity (under multiplication) in an algebraically closed field of characteristic 0 ; we then put . We leave the reader to rewrite the preceeding results for this special case in the corresponding notation.
Example 2
Let be a nonzero element of . The map from into induces an isomorphism on quotient modules from onto the submodule of consisting of elements annihilated by . If is an -module annihilated by , or equivalently an -module, then the -module is identified with . We leave the reader to rewrite the preceding results for this special case in the corresponding notation (cf. V, p. 86).
Exercises
See the exercises for § 4.
[^1]: We will see later (A, X, p. 18) that the A-module K/A is injective. It follows that Prop. 11 remains valid for arbitrary A-modules M, M' and M".