You are on page 1of 16

PRODUCT BASES AND ENDOMORPHISMS

OF PRODUCTS OF INTEGERS
Copyright c ( 2007-2009 All Rights Reserved
E. F. CORNELIUS, JR.
College of Engineering and Science, University of Detroit Mercy,
Detroit, MI, USA 48221-3038
Email: efcornelius@comcast.net, frankcornelius@alum.mit.edu
Abstract
Let C(/. N) denote the /-tuples of nonnegative integers N. The endomor-
phism ring of the group of all integer-valued functions on C(/. N) is isomorphic
to the ring of /-dimensional row-…nite in…nite matrices over the integers. The
product bases of that group are represented by the multiplicative group of in-
vertible elements in that matrix ring. In particular, all endomorphisms of the
Baer-Specker group are multiplication by row-…nite in…nite integral matrices.
All products in the Baer-Specker group are characterized, and a lemma of Lás-
zló Fuchs regarding such products is revisited.
1. Introduction
The integral root basis + = ¦j
n
(r) : : ÷ N¦ of the group of integral poly-
nomials Z[r] was introduced in [CS07] as j
0
(r) = 1, j
n+1
(r) = j
n
(r)(r ÷ :)
for : _ 0. In [CS08] this basis was extended to a basis of the group of inte-
gral polynomials in / variables, Z[A
k
], which then was used to construct the
group of formal power series with respect to the extended integral root basis,
Z[[A
k
]]
<
= ¦
P
m2C(k;N)
a
m
j
m
: a
m
÷ Z¦. The integral root basis has the
property that the valuation map · : Z[[A
k
]]
<
÷ Z
C(k;N)
de…ned by ·1(b) =
1(b), 1 ÷ Z[[A
k
]]
<
, b ÷ C(/. N), is a monomorphism, the image of which is the
group of multinomial points, '1(/. N) · Z
C(k;N)
.
A principal result of [CS08, Theorem 5.4] is that Z
C(k;N)
has a subset / =
¦c
m
: m÷ C(/. N)¦ and '1(/. N) has a subset ( = ¦
m
: m÷ C(/. N)¦ which
satisfy:
(1) For all a, m÷ C(/. N),
m
(a) = j
m
(a).
(2) Each 1 ÷ Z
C(k;N)
has a unique representation of the form 1 =
P
m2C(k;N)
/
m
c
m
: /
m
÷ Z.
(3) Each o ÷ '1(/. N) has a unique representation of the form o =
P
m2C(k;N)
c
m

m
: c
m
÷ Z.
1
(4) For all : ÷ C(/. N),
m
= j
m
c
m
where j
m
is a positive integer.
(5) Let m = (:
0
. . . . . :
k1
), n = (:
0
. . . . . :
k1
) ÷ C(/. N). If :
i
_ :
i
for 0 _ i _ / ÷1, then from (4), j
m
[ j
n
.
The structure of Z
C(k;N)
´'1(/. N) also was described in [CS08, Theorem 5.4]
and a test for inclusion in '1(/. N) was formulated in [CS08, Theorem 5.7].
From such observations evolved the de…nition of a product basis. Let l =
¦n
m
: m÷ C(/. N)¦ be a subset of Z
C(k;N)
. If each expression
P
m2C(k;N)
/
m
n
m
,
/
m
÷ Z, represents an element of Z
C(k;N)
, the set of all such elements forms a
subgroup of Z
C(k;N)
, denoted
Q
m2C(k;N)
'n
m
`, which is called a product in
Z
C(k;N)
. If the representations
P
m2C(k;N)
/
m
n
m
, /
m
÷ Z, are unique, then l is
said to be a product basis of
Q
m2C(k;N)
'n
m
`. The principal objective of this
paper is to characterize products and product bases in term of endomorphisms
of Z
C(k;N)
.
Abelian group notation and terminology are standard [F70 & 73]. N denotes
the nonnegative integers and C(/. N) the /-tuples of nonnegative integers for the
positive integer /. Elements of C(/. N) are denoted by bold lower-case letters,
m, n, etc. C(/. :) · C(/. N) denotes all /-tuples with entries < :, : ÷ N
+
,
the positive integers. Z denotes the ring of integers. The group of all functions
from C(/. N) to Z is denoted Z
C(k;N)
. Although Z
C(k;N)
is in fact a ring under
pointwise multiplication, that fact is merely noted here.
Let P = Z
C(1;N)
= Z
N
=
Q
N
Z, the group of all integer sequences, known as
the Baer-Specker group [B37, S50]. Although it is true that Z
C(k;N) ~
= P
k ~
= P,
there is a de…nite di¤erence when attempting to …t polynomials. For exam-
ple, no permutation of the sequence (1. 0. 0. 0) can be generated by an integral
polynomial in one variable, because in any such permutation (a
0
. a
1
. a
2
. a
3
), the
expression a
0
´6 ÷ a
1
´2 + a
2
´2 ÷ a
3
´6 can never be an integer, a prerequisite
for such generation. However, all arrangements of (1. 0. 0. 0) indexed by the ele-
ments of C(2. 2) can be generated by integral polynomials in 2 variables. [CS08,
Section 2]. Groups are studied from the perspective of Z
C(k;N)
for that reason,
so that this paper is a continuation of [CS08].
2. k-Dimensional In…nite Integral Matrices And Their Operations
A /-dimensional in…nite integral matrix is a function ' : C(/. N)C(/. N) ÷
Z, the value of which at (i. j) ÷ C(/. N) C(/. N) is written as '(i. j). The
group of all /-dimensional in…nite integral matrices is denoted ´
Z
(/. .), with
addition being de…ned pointwise.
1 ÷ ´
Z
(/. .) is said to be row-…nite if, for each i, 1(i. j) = 0 for all but
…nitely many j; i.e., if for each i, there exists l
i
such that 1(i. j) = 0 unless j _ l
i
2
(at each of the / coordinates). A column-…nite matrix is de…ned analogously.
The group of all row-…nite matrices in ´
Z
(/. .) is denoted T´
Z
(/. .). T ÷
´
Z
(/. .) is said to be lower-triangular if T(i. j) = 0 unless i _ j. The group
of all lower-triangular matrices in ´
Z
(/. .) is denoted T ´
Z
(/. .). Clearly
T ´
Z
(/. .) · T´
Z
(/. .).
For 1 ÷ T´
Z
(/. .) and G ÷ ´
Z
(/. .), the product of 1 and G, denoted
1G ÷ ´
Z
(/. .), is de…ned as
1G(i. j) =
P
l2C(k;N)
1(i. l)G(l. j)
Under multiplication thus de…ned, T´
Z
(/. .) and T ´
Z
(/. .) are unital rings
with identity 1(i. j) = c(i. j), the Kronecker delta.
The j
th
column of ' ÷ ´
Z
(/. .) (or column j of '), denoted '(j), is the
function C(/. N) ÷ Z de…ned by '(j)(i) = '(i. j). Given i = ¦1
j
÷ Z
C(k;N)
:
j ÷ C(/. N)¦, the matrix with columns 1
j
is de…ned to be '

÷ ´
Z
(/. .)
such that '

(i. j) = 1
j
(i) = '

(j)(i), i, j ÷ C(/. N). A column operation on
' ÷ ´
Z
(/. .) is an operation of the following type:
(a) Multiply a column of ' by ÷1; for example, multiply the l
th
column
of ' by ÷1 to obtain '
0
÷ ´
Z
(/. .) satisfying '
0
(i. j) = '(i. j) for j = l and
'
0
(i. l) = ÷'(i. l) = ÷'(l)(i).
(b) Interchange two columns of '; for example, interchange columns l
1
and l
2
of ' to obtain '
0
÷ ´
Z
(/. .) satisfying '
0
(i. j) = '(i. j) for j = l
1
and l
2
and '
0
(i. l
1
) = '(i. l
2
), '
0
(i. l
2
) = '(i. l
1
).
(c) Add an integral multiple of one column of ' to another column of ';
for example, for any integer :, add : times column l
1
of ' to column l
2
of '
to obtain '
0
÷ ´
Z
(/. .) satisfying '
0
(i. j) = '(i. j) for j = l
2
, and '
0
(i. l
2
) =
:'(i. l
1
) + '(i. l
2
) = :'(l
1
)(i) + '(l
2
)(i). Subtraction is accomplished with
÷:.
1 ÷ T´
Z
(/. .) is said to be invertible if there exists G ÷ T´
Z
(/. .)
satisfying 1G = 1 = G1. It clear that such a G is a unique two-sided inverse, so
that 1
1
= G is well de…ned [C50, pp 21-25]. The set of invertible matrices in

Z
(/. .) is denoted (L
Z
(/. .) and forms a group under matrix multiplication.
Proposition 2.1. If 1
0
÷ T´
Z
(/. .) is obtained from 1 ÷ T´
Z
(/. .)
through one of the column operations described in (a), (b) or (c) above, then 1
0
= 11 where 1 is obtained by performing the same operation on 1. 1 is both
row- and column-…nite and invertible.
3
Proof. Case (c) above; the others are far simpler. Suppose that for the
integer :, : times column l
1
of 1 has been added to column l
2
of 1 to obtain
1
0
÷ T´
Z
(/. .) satisfying 1
0
(i. j) = 1(i. j) for j = l
2
and 1
0
(i. l
2
) = :1(i. l
1
)
+ 1(i. l
2
) = :1(l
1
)(i) + 1(l
2
)(i). Let 1(i. j) = 1(i. j) for j = l
2
and 1(i. l
2
) =
:1(i. l
1
) + 1(i. l
2
) = :c(i. l
1
) + c(i. l
2
). Then 11(i. j) =
P
l2C(k;N)
1(i. l)1(l. j)
= (
P
l2C(k;N);l6=l
2
1(i. l)1(l. j)) + 1(i. l
2
)1(l
2
. j). When j = l
2
, this expression
becomes
0
@
X
l2C(k;N);l6=l
2
1(i. l)c(l. j)
1
A
+ 1(i. l
2
)c(l
2
. j) =
0
@
X
l2C(k;N);l6=l
2
1(i. l)c(l. j)
1
A
+ 0
= 1(i. j)c(j. j) = 1(i. j) = 1
0
(i. j).
When j = l
2
, the expression becomes
0
@
X
l2C(k;N);l6=l
2
1(i. l)1(l. l
2
)
1
A
+ 1(i. l
2
)1(l
2
. l
2
) =
0
@
X
l2C(k;N);l6=l
2
1(i. l)[:c(l. l
1
) + c(l. l
2
)]
1
A
+ 1(i. l
2
)[:c(l
2
. l
1
) + c(l
2
. l
2
)] =
1(i. l
1
):c(l
1
. l
1
)+ 1(i. l
2
) = :1(i. l
1
) + 1(i. l
2
) = 1
0
(i. l
2
).
It is clear that the inverse of 1 is the matrix which subtracts : times column
l
1
of 1 from column l
2
of 1. 1 and its inverse obviously are row- and column-
…nite.
The corollary is immediate:
Corollary 2.2. If 1
0
÷ T´
Z
(/. .) is obtained from 1 ÷ T´
Z
(/. .)
through a …nite sequence of operations described in (a), (b) or (c) above, then
1
0
= 11 where 1 is obtained by performing the same sequence of operations
on 1. 1 is both row- and column-…nite and invertible.
For 1 ÷ T´
Z
(/. .) and 1 ÷ Z
C(k;N)
, the matrix-vector product of 1 and 1,
denoted simply 11, is de…ned as 11 : C(/. N) ÷Z, 11(i) =
P
l2C(k;N)
1(i. l)1(l).
Each such 1 induces an endomorphism of Z
C(k;N)
. A product in Z
C(k;N)
,
Q
m2C(k;N)
'n
m
`, is simply the image of such an endomorphism where the matrix
1 is constructed with columns n
m
, m÷ C(/. N).
Some of the results below can be proved through topological techniques.
Background material may be found in [BI01, BI96(1) & (2)] and the references
cited therein. For 1 = o ÷ Z
C(k;N)
, the distance between them, d(1. o), is de…ned
to be 2
n
, where : is the least positive integer for which there is an x ÷ C(/. :)
with 1(x) = o(x). d(1. 1) = 0, of course. It is easy to check that d is a metric
on Z
C(k;N)
, which comports with the product topology on Z
C(k;N)
, when Z is
discrete.
4
Z
C(k;N)
is a separable metric space having as a countable dense subset those
functions 1 which are 0 outside some C(/. :
f
), :
f
÷ N
+
. This countable
dense subset also is a free subgroup, which is denoted by (/); (1) = ·
P, the sequences of integers which are 0 after a while. (/) has a standard
basis, ¦c
j
(i) = c(i. j) : i, j ÷C(/. N)¦, which also the standard product basis for
Z
C(k;N)
. Finally, Z
C(k;N)
is a complete metric space in which Cauchy sequences
eventually become constant pointwise.
The important aspects of this topology are that all endomorphisms of Z
C(k;N)
are continuous and are determined by their values on (/) and hence by their
values on the standard basis. As a result, with each element 1 in the endo-
morphism ring of Z
C(k;N)
, 1:d(Z
C(k;N)
), is associated a unique element '
f
÷
´
Z
(/. .), given by '
f
(i. j) = [1(c
j
)](i), i, j ÷ C(/. N). Recall that each 1 ÷

Z
(/. .) acts on Z
C(k;N)
to produce an endomorphism of Z
C(k;N)
,
(1o)(i) =
P
l2C(k;N)
1(i. l)o(l) for o ÷ Z
C(k;N)
.
Because 1 is row-…nite, the sum reduces to
(1o)(i) =
P
l2C(k;n
i
)
1(i. l)o(l) where :
i
÷ N
+
.
The immediate goal is to prove the converse—that every endomorphism of
Z
C(k;N)
takes this form.
3. Product Bases And Endomorphisms
Analyses of the endomorphisms of Z
C(k;N)
demonstrate that its products
all stem from T ´
Z
(/. .) and its product bases from (L
Z
(/. .). Of particular
interest is the fact that every endomorphism of P is nothing more than multi-
plication by a row-…nite matrix, and every product basis of P is produced by
an integrally invertible matrix.
Theorem 3.1. Every endomorphism of Z
C(k;N)
is induced by the action
of a /-dimensional row-…nite in…nite integral matrix.
Proof. Let 1 ÷ 1:d(Z
C(k;N)
), and let '
f
÷ ´
Z
(/. .) be de…ned by '
f
(i. j)
= [1(c
j
)](i), i, j ÷ C(/. N). Suppose that '
f
is not row-…nite; i.e., suppose there
are i and distinct j
i
÷ C(/. N) such that '
f
(i. j
i
) = 0, i ÷ N. De…ne o ÷ Z
C(k;N)
by o(m) = '
f
(i. j
i
) if m = j
i
, i ÷ N, and o(m) = 0 otherwise, for all m÷
C(/. N). Further, for : ÷ N
+
, de…ne o
n
(m) = o(m) for m÷ C(/. :) and o
n
(m)
= 0 for m ÷ C(/. N)`C(/. :). By design, o
n
converges to o, so that 1(o
n
)
must converge to 1(o). Now o
n
=
P
l2C(k;n)
o
n
(l)c
l
=
P
ji2C(k;n)
o
n
(j
i
)c
ji
=
P
ji2C(k;n)
'
f
(i. j
i
)c
ji
so that 1(o
n
) =
P
ji2C(k;n)
'
f
(i. j
i
)1(c
ji
) and
[1(o
n
)](i) =
X
ji2C(k;n)
'
f
(i. j
i
)[1(c
ji
)](i) =
5
X
ji2C(k;n)
'
f
(i. j
i
)'
f
(i. j
i
) =
X
ji2C(k;n)
'
f
(i. j
i
)
2
.
Now [1(o)](i) = lim
n!1
[1(o
n
)](i) = lim
n!1
P
ji2C(k;n)
'
f
(i. j
i
)
2
. Since
[1(o)](i) ÷ Z, it is not possible that an in…nite number of the '
f
(i. j
i
) are
nonzero; i.e., '
f
must be row-…nite.
To see that '
f
acts on Z
C(k;N)
to produce 1, it su¢ces to check agreement on
the standard basis of (/). For i, j ÷ C(/. N), ('
f
c
j
)(i) =
P
l2C(k;N)
'
f
(i. l)c
j
(l)
=
P
l2C(k;N)
[1(c
l
)](i)c(j. l) = [1(c
j
)](i).
Corollary 3.2. (i) The endomorphism ring of Z
C(k;N)
, 1:d(Z
C(k;N)
), is
isomorphic to the ring of /-dimensional row-…nite in…nite integral matrices,

Z
(/. .).
(ii) The automorphism group of Z
C(k;N)
, ¹nt(Z
C(k;N)
), is isomorphic to
(L
Z
(/. .), the multiplicative group of invertible matrices in T´
Z
(/. .).
Proof. For 1 ÷ 1:d(Z
C(k;N)
), 1 ÷ '1 is a ring isomorphism with

Z
(/. .), which maps ¹nt(Z
C(k;N)
) isomorphically onto (L
Z
(/. .). The row-
…niteness of the elements of T´
Z
(/. .) ensures that all sums are …nite and that
matrix multiplication is associative.
Corollary 3.3. There is a one-to-one correspondence between the matrices
of (L
Z
(/. .) and the product bases of Z
C(k;N)
.
Proof. Let ¦n
m
: m÷ C(/. N)¦ be a product basis for Z
C(k;N)
, and let
¹ ÷ ¹nt(Z
C(k;N)
) be de…ned by [¹(1)](m) = a
m
for 1 ÷ Z
C(k;N)
, when 1 =
P
m2C(k;N)
a
m
n
m
, a
m
÷ Z. The uniqueness of expression of the elements of
Z
C(k;N)
in terms of the product basis guarantees that ¹ is an automorphism.
By Corollary 3.2(ii), ¹ corresponds to a unique element of (L
Z
(/. .). The
converse is clear.
Corollary 3.4. If l is a subgroup of Z
C(k;N)
with product basis l =
Q
m2C(k;N)
'n
m
` and j = ¦n
m
: m÷ C(/. N)¦, then the matrix with columns
n
m
, '

, is row-…nite and l = '

Z
C(k;N)
.
Proof. The mapping + : Z
C(k;N)
÷ l _ Z
C(k;N)
de…ned for 1 ÷ Z
C(k;N)
by
1 ÷
P
m2C(k;N)
1(m)n
m
is an endomorphism of Z
C(k;N)
, so according to The-
orem 3.1, it is induced by a row-…nite matrix '

satisfying '

(i. j) = [+(c
j
)](i)
for all i, j ÷ C(/. N), and '

Z
C(k;N)
= l. Now +(c
j
) =
P
m2C(k;N)
c
j
(m)n
m
= n
j
, so that [+(c
j
)](i) = n
j
(i) = '

(i. j) for all i, j ÷ C(/. N), demonstrating
that '

= '

.
6
4. All Products In Z
C(k;N)
Are Generated By Lower-Triangular Matrices
Recall that when / = 1, C(/. N) reduces to N and Z
C(k;N)
reduces to Z
N
,
the Baer-Specker group P. Further, T´
Z
(1. .) reduces to the familiar ring of
2-dimensional row-…nite in…nite integral matrices. With this understanding, a
product in P is just a product in the sense of Fuchs [F73, p 164].
Lemma 4.1. Every product in P is generated by a lower-triangular matrix,
the columns of which form a basis of the product.
Proof. Let 1 ÷ T´
Z
(1. .). The proof proceeds by transforming 1 via
column operations into a lower-triangular matrix T satisfying TP = 1P. Since
1 = 0 is trivial, assume 1 = 0. Let n ÷ 1P, 0 = n = (n(0). n(1). ...)
t
, where t
denotes transpose, and suppose n = 1r for r = (r(0). r(1). ...)
t
÷ P.
If row 0 of 1 consists entirely of 0’s, simply set 1
1
= 1, 1
1
= 1, and r
1
=
r and proceed to the next step. Otherwise, perform column operations on the
nonzero elements of row 0 of 1 to produce a new matrix 1
1
in which 1
1
(0. ,) =
0 for all , 0. This is tantamount to …nding the gcd of the elements in row 0,
columns , _ 0 of 1. Perform the same column operations on 1 to produce 1
1
,
so that 1
1
= 11
1
with 1
1
an invertible row- and column-…nite matrix. Since 1
= 1
1
1
1
1
, n = 1r = 1
1
(1
1
1
r). Set r
1
= (r
1
(0). r
1
(1). ...)
t
= 1
1
1
r so that n =
1
1
r
1
.
If 1
1
(1. ,) = 0 for all , _ 1, simply set 1
2
= 1
1
, r
2
= r
1
, and 1
2
= 1 and
proceed. Otherwise perform column operations on the nonzero elements of row
1 in columns , _ 1 of 1
1
to produce a new matrix 1
2
in which 1
2
(1. ,) = 0 for
all , 1. As before, this is tantamount to …nding the gcd of the elements in
row 1, columns , _ 1 of 1
1
. 1
2
(i. 0) = 1
1
(i. 0) for all i; i.e., column 0 of 1
2
is
the same as column 0 of 1
1
.
Perform the same column operations on 1 to produce 1
2
, so that 1
2
=
1
1
1
2
with 1
2
an invertible row- and column-…nite matrix. Note that since no
operation was performed on column 0 of 1
1
, none was performed on column 0
of 1. Moreover rows 0 of 1
1
and 1 remained unchanged (because of 0 entries),
so that 1
2
(0. 0) = 1 and the rest of the row 0 and column 0 entries of 1
2
are 0.
Thus the same is true of 1
1
2
. Since 1 = 1
2
1
1
2
1
1
1
, n = 1r = 1
2
1
1
2
(1
1
1
r
0
) =
1
2
(1
1
2
r
1
). Set r
2
= 1
1
2
r
1
so that n = 1
2
r
2
. Because row 0 of 1
1
2
is 1. 0. . . .,
it follows that r
2
(0) = r
1
(0).
Suppose that after : 1 iterations, row-…nite matrices 1
1
, . . . , 1
n
and
invertible row- and column-…nite matrices 1
1
, . . . , 1
n
and r
1
, . . . , r
n
÷ P have
been obtained, such that
7
(a) 1
n
(i. ,) = 0 for all , i, 0 _ i _ : ÷ 1; i.e., the matrices 1
n
are
becoming increasingly lower-triangular.
(b) 1
n
(,) = 1
n1
(,) for 0 _ , < : ÷1; i.e., after the :
th
iteration, rows
and columns < : do not change.
(c) 1
n
= 1
n1
1
n
= 11
1
1
n
.
(d) 1
n
(i. ,) = c(i. ,) = 1
1
n
(i. ,), 0 _ i < :÷1, , ÷ N; and 0 _ , < :÷1,
i ÷ N; i.e., the 1
n
’s and their inverses are becoming increasingly diagonal with
1’s down the diagonal, tending toward the identity matrix.
(e) r
n
= 1
1
n
r
n1
, n = 1
n
r
n
; i.e., n is always in the image of 1
n
.
(f) r
n
(i) = r
n1
(i), 0 _ i < : ÷1; i.e., after the :
th
iteration, the row
elements < : of the r
n
’s do not change.
If 1
n
(:. ,) = 0 for all , _ :, simply set 1
n+1
= 1
n
, 1
n+1
= 1, and r
n+1
=
r
n
. Otherwise, perform column operations on the nonzero elements of the :
th
row in columns , _ : of 1
n
to produce a new matrix 1
n+1
in which 1
n+1
(:. ,)
= 0 for all , :. Because no operations are performed on columns 0-(:-1),
1
n+1
(i. ,) = 1
n
(i. ,), i ÷ N, , = 0, . . . , : ÷1; i.e., columns 0-(:-1) of 1
n+1
are
the same as columns 0-(:-1) of 1
n
.
Perform the same column operations on 1 to produce 1
n+1
, so that 1
n+1
=
1
n
1
n+1
with 1
n+1
an invertible row- and column-…nite matrix. Note that since
no operation was performed on columns 0-(:-1) of 1
n
, none was performed
on columns 0-(:-1) of 1. Moreover, rows 0-(:-1) of 1
n
were not changed (be-
cause of 0’s), and the same is true of 1 and 1
1
n+1
. (i) As a result, 1
n+1
(i. ,)
= c(i. ,) = 1
1
n+1
(i. ,), i < :, , ÷ N, and , < :, i ÷ N. Since 1
n+1
=
1
n
1
n+1
= (11
1
1
n
)1
n+1
, 1 = 1
n+1
1
1
n+1
1
1
1
, n = 1r = 1
n+1
1
1
n+1
1
1
1
r
= 1
n+1
(1
1
n+1
r
n
). (ii) Set r
n+1
= 1
1
n+1
r
n
so that n = 1
n+1
r
n+1
. (i) and (ii)
imply that r
n+1
(i) = r
n
(i), i = 0, . . . , : ÷1.
The Cauchy sequence r
n
converges to some r
0
÷ P because P is a complete
metric space. Since all endomorphisms of P are continuous, the sequence 1
n
(r
j
)
converges to 1
n
(r
0
) for each :. Because of (b) and (f), the convergence is
uniform. The Cauchy sequence of matrices 1
n
thus obtained converges to a
lower-triangular matrix T and Tr
0
= n so that TP _ 1P. From (b) it follows
that TP _ 1P and so TP = 1P.
It may, of course, happen that some of the diagonal terms of T are 0 so that
the …rst nonzero entry in a column may be below the diagonal. Should that
occur, the …rst nonzero entry in the next column will be in a lower row. If T
has only …nitely many nonzero columns, the result is elementary. If all columns
of T are nonzero, it is clear that an in…nite linear combination of its columns
can be 0 only if all coe¢cients are 0, so that the columns of T form a product
basis.
Remark 4.2. From the results in Section 3, a row-…nite matrix can be
classi…ed as epic or monic if the endomorphism which it induces is epic or monic,
8
respectively. (1) If 1 ÷ T´
Z
(1. .) has a 0 row :, it cannot be epic because
c
m
cannot be reached. If it has a 0 column :, it cannot be monic because c
n
is in
the kernel. Here 1 = ¦c
i
: i ÷ N¦ denote the usual basis of · P; 1 is also the
standard product basis of P. (2) Similarly, if a row (column) of 1 is a multiple
of another row (column), then 1 cannot be epic (monic). (3) T ÷ T ´
Z
(1. .)
is invertible (2-sided) over the rationals if and only if each diagonal entry is
nonzero [C50, pp 19-20 ]. T is integrally invertible if and only if each diagonal
entry = ±1.
Lemma 4.3. T ÷ T ´
Z
(1. .) is epic if and only if each diagonal entry is
±1.
Proof. If each diagonal entry of T is ±1, then T ÷ (L
Z
(1. .) and so is epic.
Conversely let n = (n(0). n(1). ...)
t
be a nonzero element of P, where t denotes
transpose. To solve Tr = T(r(0). r(1). ...)
t
= n, induct on :. Let T
n
= (T[i. ,]),
i, , = 0, 1, . . . , :. If : = 0 and T(0. 0) = ±1, then T(0. 0)r(0) = n(0) is not
solvable in general. If T(0. 0) = ±1, then r(0) is uniquely determined. Suppose
that the diagonals T(:. :) = ±1 for : = 0, 1, . . . , :, and suppose further
that T
n
(r(0). .... r(:))
t
= (n(0). .... n(:))
t
is uniquely solvable, (r(0). .... r(:))
t
= T
1
n
(n(0). .... n(:))
t
. Express T
n+1
=

T
n
0
T(: + 1. 0)... T(: + 1. : + 1)

so
that T
n+1
(r(0). .... r(:). r(: + 1))
t
=

T
n
(r(0). .... r(:))
t
P
j=n+1
j=0
T(: + 1. ,)r(,)

, with some
abuse of notation for the sake of simplicity. Now by the induction hypothe-
sis, (r(0). .... r(:))
t
= T
1
n
(n(0). .... n(:))
t
uniquely. Thus the only solution to
P
j=n+1
j=0
T(: + 1. ,)r(,) = n(: + 1) must come from T(: + 1. : + 1)r(: + 1) =
n(: + 1) –
P
j=n
j=0
T(: + 1. ,)r(,), which is impossible in general unless T(: +
1. : + 1) = ±1.
Invertible lower-triangular in…nite integral matrices may therefore be char-
acterized as follows:
Theorem 4.4. For T ÷ T ´
Z
(1. .) the following are equivalent:
(1) T is epic.
(2) T(i. i) = ±1 for all i ÷ N.
(3) T ÷ (L
Z
(1. .).
The following lemma is a generalization of Lemma 4.1.
Lemma 4.5. Every 1 ÷ T´
Z
(/. .) can be triangularized in the sense
that there exists 1
0
÷ T´
Z
(/. .) satisfying 1
0
[(i. i). (j. ,)] = 0 for all i. j ÷
C(/ ÷1. N). i < , ÷ N, and 1
0
Z
C(k;N)
= 1Z
C(k;N)
.
Proof. De…ne 1
ij
÷ T´
Z
(1. .) for i. j ÷ C(/ ÷ 1. N) and i. , ÷ N by
1
ij
(i. ,) = 1[(i. i). (j. ,)]. As in the proof of Lemma 4.1, there exists T
ij
÷
9
T ´
Z
(1. .), obtained by performing column operations on the 2-dimensional
matrix 1
ij
, which satis…es T
ij
P = 1
ij
P. For i. j ÷ C(/ ÷ 1. N) and i. , ÷ N,
de…ne 1
0
[(i. i). (j. ,)] = T
ij
(i. ,). Since each T
ij
is lower-triangular, it is clear
that 1
0
[(i. i). (j. ,)] = 0 for all i < ,.
To see that 1
0
is row-…nite, …x (i. i). T
ij
(i. ,) = 0 for all , i; i.e., T
ij
(i. ,) =
0 for only …nitely many ,, namely , = 0. . . . . i. For each such ,, there are only
…nitely many j where 1[(i. i). (j. ,)] = 0. Thus, since the entries of each T
ij
were obtained from those of 1, for each (i. i), only …nitely many T
ij
(i. ,) = 0,
so 1
0
÷ T´
Z
(/. .).
Finally, to see that 1
0
Z
C(k;N)
= 1Z
C(k;N)
, let 1 ÷ Z
C(k;N)
. For i ÷ C(/÷1. N)
and i ÷ N, (11)(i. i) =
P
j2C(k;N)
1[(i. i). j]1(j) =
P
j2C(k1;N)
P
j2N
1[(i. i). (j. ,)]1(j. ,)
=
P
j2C(k1;N)
P
j2N
1
ij
(i. ,)n
j
(,), with n
j
÷ P de…ned by n
j
(,) = 1(j. ,) for j ÷ C(/÷
1. N) and , ÷ N. Since T
ij
P = 1
ij
P, there exists r
j
÷ P satisfying T
ij
r
j
= 1
ij
n
j
or, for i ÷ N, (T
ij
r
j
)(i) = (1
ij
n
j
)(i); i.e.,
P
j2N
T
ij
(i. ,)r
j
(,) =
P
j2N
1
ij
(i. ,)n
j
(,),
so that (11)(i. i) =
P
j2C(k1;N)
P
j2N
1
ij
(i. ,)n
j
(,) =
P
j2C(k1;N)
P
j2N
T
ij
(i. ,)r
j
(,) =
P
j2C(k1;N)
P
j2N
1
0
[(i. i). (j. ,)]r
j
(,) =
P
j2C(k1;N)
P
j2N
1
0
[(i. i). (j. ,)]o(j. ,) = (1
0
o)(i. i) when o ÷ Z
C(k;N)
is de…ned as
o(j. ,) = r
j
(,) for j ÷ C(/ ÷1. N) and , ÷ N.
In the Proof of Lemma 4.5, notation has been abused for the sake of sim-
plicity, but attention to each index set should clear up any confusion.
A principal result is the following
Theorem 4.6. Every product in Z
C(k;N)
is generated by a lower-triangular
matrix, the columns of which form a basis of the product.
Proof. By induction on /. Lemma 4.1 establishes the case / = 1, so the
result holds for some /. Let 1 ÷ T´
Z
(/ + 1. .) so that 1Z
C(k+1;N)
is a
product in Z
C(k+1;N)
. 1 = 0 is trivial so assume that 1 = 0. Per Lemma
4.5, triangularize 1; i.e., …nd 1
0
÷ T´
Z
(/ + 1. .) such that 1
0
[(i. i). (j. ,)] =
0 for all i, j ÷ C(/. N), i < , ÷ N, and such that 1
0
Z
C(k+1;N)
= 1Z
C(k+1;N)
.
When i = 0, by the induction hypothesis, there exists T
00
÷ T ´
Z
(/. .) such
that T
00
Z
C(k;N)
= 1
0
[(i. 0). (j. 0)]Z
C(k;N)
, where by 1
0
[(i. 0). (j. 0)] is meant the
matrix G
0
÷ T´
Z
(/. .), de…ned as G
0
(i. j) = 1
0
[(i. 0). (j. 0)] and similarly
below; again notation is abused for the sake of simplicity. De…ne 1
1
÷ T´
Z
(/+
1. .) as follows:
For all i. j ÷ C(/. N) and i. , ÷ N.
1
1
[(i. i). (j. ,)] =

T
00
(i. j) if i = 0 = ,
1
0
[(i. i). (j. ,)] otherwise
10
Observe that 1
1
[(i. 0). (j. 0)] = 0 unless i _ j, and that 1
1
Z
C(k+1;N)
= 1
0
Z
C(k+1;N)
= 1Z
C(k+1;N)
.
When i = 1, by the induction hypothesis, there exist T
10
and T
11
÷ T ´
Z
(/. .)
such that
T
10
Z
C(k;N)
= 1
1
[(i. 1). (j. 0)]Z
C(k;N)
and T
11
Z
C(k;N)
= 1
1
[(i. 1). (j. 1)]Z
C(k;N)
.
De…ne 1
2
÷ T´
Z
(/ + 1. .) by:
For all i. j ÷ C(/. N) and i. , ÷ N.
1
2
[(i. i). (j. ,)] =
8
<
:
T
10
(i. j) if i = 1 and , = 0
T
11
(i. j) if i = 1 and , = 1
1
1
[(i. i). (j. ,)] otherwise
Observe that 0 = 1
2
[(i. 0). (j. 0)] = 1
2
[(i. 1). (j. 0)] = 1
2
[(i. 1). (j. 1)] unless i _ j,
and that 1
2
Z
C(k+1;N)
= 1
1
Z
C(k+1;N)
= 1
0
Z
C(k+1;N)
= 1Z
C(k+1;N)
.
Proceeding in this fashion, obtain 1
n
÷ T´
Z
(/+1. .) and T
ij
÷ T ´
Z
(/. .)
with these properties: For i, j ÷ C(/. N) and i, , ÷ N,
(a) T
ij
Z
C(k;N)
= 1
n1
[(i. i). (j. ,)]Z
C(k;N)
, : _ i _ ,;
(b) 1
n
[(i. i). (j. ,)] = 0 unless (i. i) _ (j. ,) for i, , _ :; and
(c) 1
n
Z
C(k+1;N)
= 1Z
C(k+1;N)
.
Then as in Lemma 4.1, lim
n!1
1
n
= T uniformly and TZ
C(k+1;N)
= 1Z
C(k+1;N)
with T ÷ T ´
Z
(/ + 1. .).
By use of the foregoing methodology, the following readily can be proved:
Proposition 4.7. Let T ÷ T ´
Z
(/. .).
(1) These are equivalent: (i) T is epic; (ii) T(i. i) = ±1 for all i ÷
C(/. N); (iii) T ÷ (L
Z
(/. .).
(2) T is monic if and only if T(i. i) = 0 for all i ÷ C(/. N).
11
5. In…nite Matrix Examples And Counterexamples
The next examples and counterexamples illustrate results from Section 4 and
further illustrate some of the pitfalls and limitations of working with in…nite
matrices. They are not new but continue to be discussed in various contexts.
[BM05, sec 2]; [CC-CS01, p 437]; [BI96(1), pp 189-191].
As before, the elements of P are viewed as column vectors and 1 = ¦c
j
: , ÷
N¦ denotes the usual basis of · P; 1 is also the standard product basis of
P. The same symbol 1 is used for the matrix with the c
j
as columns, i.e., the
identity matrix. The same liberty is taken with other bases and their matrices.
Example 5.1. 1
1
= ¦c
0
+ c
1
+ . . .+ c
i
: i ÷ N¦ also is a basis of ,
having matrix
1
1
=
0
B
B
B
B
B
@
1 1 1 1 ...
0 1 1 1 ...
0 0 1 1 ...
0 0 0 1 ...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
C
C
C
C
C
A
Because the matrix is not row-…nite, 1
1
cannot be a product basis of P.
Example 5.2. 1
2
= ¦c
0
, c
i
– c
i1
: i ÷ N
+
¦ also is a basis of , having
matrix
1
2
=
0
B
B
B
B
B
@
1 ÷1 0 0 ...
0 1 ÷1 0 ...
0 0 1 ÷1 ...
0 0 0 1 ...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
C
C
C
C
C
A
Note that 1
2
is not lower triangular.
Example 5.3. Let 1 ÷ P denote the column vector having all entries =
1 and let 1
3
= ¦c
0
+ c
2
+ . . .+ c
i1
–1 : i ÷ N
+
¦. Note that 1
3
is not a basis
of ; indeed, no element is even contained in . The accompanying matrix is
12
1
3
=
0
B
B
B
B
B
@
0 0 0 0 ...
÷1 0 0 0 ...
÷1 ÷1 0 0 ...
÷1 ÷1 ÷1 0 ...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
C
C
C
C
C
A
The following relationships hold:
1
1
1
2
= 1 = 1
2
1
1
1
2
1
3
= 1
Observe that although both 1
2
and 1
3
are row-…nite and even pure independent,
1
2
is not monic (1 is in the kernel), and 1
3
is not epic (c
0
is not attained). In
particular, (1
1
1
2
)1 = 11 = 1, but 1
1
(1
2
1) = 1
1
(0) = 0, so that multiplication
is not associative. Despite the fact that 1
2
P = P, 1
2
nevertheless cannot serve
as a product basis for P. The problem, of course, is that 1
2
is not invertible
in T´
Z
(/. .). Restriction to multiplicatively associative matrix rings helps
eliminate much of this in…nite matrix pathology [C50, pp 19-22].
In the proof of Lemma 4.1, although the columns of the resulting lower-
triangular matrix T obviously are independent, T nevertheless may be singular,
as illustrated by 1
3
above. Moreover, it may not be the case that that the
products 1
1
1
n
in that proof will converge to a row-…nite matrix, as further
illustrated by 1
2
. Even though 1
2
itself is not monic, the resulting lower-
triangular matrix 1, which results from column operations on 1
2
, is monic.
6. Fuchs’ Lemma 95.1 Redux
Products in P were introduced in [F73] by beginning with a countable subset
¦r
n
: : ÷ N¦ · P, with no condition other than that all sums of the form r
=
P
1
n=0
:
n
r
n
be well de…ned. This, of course, means that the in…nite matrix
with columns r
n
must be row-…nite. Lemma 95.1 in [F73] then was stated as
follows:
Let A be a product in P. There are elements a
n
÷ P and integers
/
n
(: ÷ N) such that P =
Q
1
n=0
'a
n
` and A =
Q
1
n=0
'/
n
a
n
` where
(«) /
n
[ /
n+i
if /
n
= 0 for all :, i _ 0.
The goal was to establish a product analog of the well known result for stacked
bases of free groups of …nite rank [F70, Lemma 15.4]. However, a counterexam-
ple has been produced [G81].
Perhaps another approach is to utilize the weaker de…nition of stacked bases
found in [CG70], which does not require divisibility as in («). From Lemma
13
4.1, there would be no loss of generality in assuming that the r
n
÷ P form a
lower-triangular matrix.
Lemma 695.1. Let
Q
1
n=0
'r
n
` be a product of in…nite rank in P, let o
=
P
1
n=0
'r
n
` , and let o

be the pure subgroup which the r
n
generate. Then
there exist a product basis of P =
Q
1
n=0
'a
n
` and positive integers /
n
(: ÷ N)
such that
Q
1
n=0
'/
n
a
n
` =
Q
1
n=0
'r
n
` if and only if there exist stacked bases of
o and o

such that the matrix 1

formed by the basis of o

is invertible and
the matrix 1 formed by the basis of o satis…es 1P =
Q
1
n=0
'r
n
`.
Proof. If ¦a
n
: : ÷ N¦ is a product basis of P, then the a
n
’s are pure
independent. If ¦/
n
: : ÷ N)¦ are positive integers, then
Q
1
n=0
'/
n
a
n
` is a
product in P and if o =
1
n=0
'/
n
a
n
`, then o

=
1
n=0
'a
n
` so that ¦/
n
a
n
:
: ÷ N¦ and ¦a
n
: : ÷ N¦ are stacked bases for the respective subgroups of P.
The matrix formed by the a
n
’s is in (L
Z
(1. .) by Corollary 3.4; the matrix 1
formed by the r
n
= /
n
a
n
certainly satis…es 1P =
Q
1
n=0
'r
n
`.
Conversely, let
Q
1
n=0
'r
n
` be a product in P, let o be the subgroup of P
which the r
n
’s generate, and let o

be the pure subgroup which they generate.
Suppose there are stacked bases ¦/
n
a
n
: /
n
, : ÷ N¦ of o and ¦a
n
: : ÷ N¦
of o

; because o

´o is torsion, all /
n
’s must be positive. Suppose further that
the matrix formed by the a
n
’s is invertible; then the a
n
’s are a product basis of
P and
Q
1
n=0
'/
n
a
n
` is a product in P. Finally, if the matrix 1 formed by the
/
n
a
n
’s satis…es 1P =
Q
1
n=0
'r
n
`, the proof is concluded.
Remark 6.2. (1) If
Q
1
n=0
'r
n
` is a product in P and ¦a
n
: : ÷ N¦ is
a basis of the subgroup generated by the r
n
’s, the matrix formed by the a
n
’s
may not be row-…nite. See examples 1 and 1
1
in Section 5 above. (2) In
Section 5, although all positive multiples of the basis elements in example 1
2
will yield subgroups of P with stacked bases and even though 1
2
P = P, 1
2
cannot serve as a product basis for P. (3) As previously noted, the lower-
triangular matrix T, obtained from the column operations in Lemma 4.1, may
be singular. However, Nunke [BI96, p 199] has shown that every endomorphic
image of P is either free of …nite rank or isomorphic to P itself. In the latter
case, there then exists monic 1 ÷ T´
Z
(/. .) satisfying 1P = TP. Column
operations will, of course, reveal matrix rank, …nite or in…nite.
All products in P may therefore be characterized as follows:
Theorem 6.3. Every product in P is generated by a lower-triangular
in…nite integral matrix, the rank of which determines whether the product is
free of …nite rank or isomorphic to P.
7. Obiter Dicta
14
Although the methods employed and results ontained in this paper and [CS07
& 08] could be extended to more complex structures, e.g., to …nite and countably
in…nite sequences of polynomials in a countably in…nite number of variables,
and to some larger cardinals/ordinals, e.g., along the lines of the earlier works
of Zeeman [F73, Corollary 94.6] (see also [F70, Theorem 43.2]) and the more
recent ones of Eda [E83(1) & E83(2)], the author does not envision that such
extensions will render results signi…cantly di¤erent from those already attained.
References
[B37] Baer, R., Abelian groups without elements of …nite order, Duke Math. J.
3 (1937) 68-122
[BI96(1)] Blass, A. & Irwin, J., On the group of eventually divisible integer
sequences. In: Abelian groups and modules, Arnold, D. M. & Rangaswamy, K.
M., eds., Lecture Notes in Pure and Applied Mathematics, Marcel Dekker, Inc.,
182 (1996) 181-192
[BI96(2)] Blass, A. & Irwin, J., Baer meets Baire: applications of category
arguments and descriptive set theory to Z
@0
. In: Abelian groups and modules,
Arnold, D. M. & Rangaswamy, K. M., eds., Lecture Notes in Pure and Applied
Mathematics, Marcel Dekker, Inc., 182 (1996) 193-202
[BI01] Blass, A. & Irwin, J., Free subgroups of the Baer-Specker group, Comm.
Algebra 29(12) (2001) 5769-5794
[BM05] Bartosiewicz, Z. & Mozyrska, D., Observability of row–…nite countable
systems of linear di¤erential equations, http://katmet.pb.bialystok.pl/mat.barz/
Infsys4.pdf (2005)
[C50] Cooke, R. G., In…nite Matrices and Sequence Spaces, McMillan, London
(1950)
[CC-CS01] Camillo, V.; Costa-Cano, F. J. & Simon, J. J., Relating properties
of a ring and its ring of row and column …nite matrices, J. Algebra 244 (2001)
435-449
[CG70] Cohen, J. M. & Gluck, H., Stacked bases for modules over principal
ideal domains, J. Algebra 14 (1970) 493-505
[CS07] Cornelius, E. F., Jr. & Schultz, P., Polynomial points, J. Integer Se-
quences 10 (2007) Article 07.3.6
15
[CS08] Cornelius, E. F., Jr. & Schultz, P., Multinomial points, Houston J.
Math. 34(3) (2008) 661-676
[E83(1)] Eda, K., On a Boolean power of a torsion free abelian group, J. Algebra
82(1) (1983) 84-93
[E83(2)] Eda, K., Almost-slender groups and Fuchs-44 groups, Comment. Math.
Uni. St. Paui., 32(2) (1983) 131-135
[F70] Fuchs, L., I In…nite Abelian Groups, Academic Press, New York & London
(1970)
[F73] Fuchs, L., II In…nite Abelian Groups, Academic Press, New York &
London (1973)
[G81] Goldsmith, B., A note on products of in…nite cyclic groups, Rend. Semin.
Mat. Univ. Padova, 64 (1981) 243-246
[S50] Specker, E., Additive gruppen von folgen ganzer Zahlen, Portugaliae
Math. 9 (1950) 131-140
16