The proof of our next theorem is inspired by the LöwenheimSkolem argument. We first state and prove an observation mentioned by Brendan Goldsmith. This will be used to reduce the theorem to a result in [
2]
.
Observation 3.1
Every subring of a principal ideal domain
$R$
is a subring of principal ideal domain contained in
$R$
and of the same cardinality.
Proof. Let
${R}_{1}\subseteq R$
be the given rings, let
$Q$
be the field of fractions of
${R}_{1}$
and set
${R}_{2}=R\cap Q$
which is a subring of
$R$
of cardinality
$\left{R}_{1}\right$
containing
${R}_{1}$
. We will show that
${R}_{2}$
is a principal ideal domain. If
$I$
is an ideal of
${R}_{2}$
, then
$IR$
is an ideal of
$R$
and so
$IR=aR$
for some
$a\in R$
. However
$a\in IR$
and so
$a=ir$
for some
$i\in I,r\in R$
. Now if
$x$
is an arbitrary element of
$I$
, then
$x=a{r}_{x}$
for some
${r}_{x}\in R$
and so
$x=ir{r}_{x}=i{t}_{x}$
say where
${t}_{x}=r{r}_{x}\in R$
. However
${t}_{x}=x/i\in K\cap R={R}_{2}$
. So
$x=i{t}_{x}$
is a product of an element in
$I$
and in
${R}_{2}$
. Since
$x$
was arbitrary in
$I$
we have
$I\subseteq i{R}_{2}$
and since the reverse inclusion is trivial, we deduce that
$I=i{R}_{2}$
is principal.
Theorem 3.2
(ZFC + MA) Let
$R$
be a principal ideal domain. Any strongly
${\aleph}_{1}$
free
$R$
module of rank
${\aleph}_{1}$
is a Crawley module.
Proof. Let
${N}_{1},{N}_{2}\subseteq G$
be two pure and dense submodules of corank
$1$
of the strongly
${\aleph}_{1}$
free
$R$
module
$G$
of rank
${\aleph}_{1}$
. Moreover let
${\phi}_{i}:G\u27f6Q$
be the canonical epimorphisms with kernel
$Ker{\phi}_{i}={N}_{i}(i=1,2)$
. Choose an
${\aleph}_{1}$
filtration
$G={\cup}_{\alpha \in {\omega}_{1}}{G}_{\alpha}$
and let
${N}_{i\alpha}={N}_{i}\cap {G}_{\alpha}$
be the induced
${\aleph}_{1}$
filtration on
${N}_{i}(i=1,2)$
. We can assume that there are
${z}_{i}\in {G}_{0}$
with
${z}_{i}{\phi}_{i}=1\in Q$
for
$i=1,2$
. Moreover, let each
${G}_{0},{N}_{i0}$
have infinite rank (w.l.o.g.). Also choose an
$R$
basis for each
${G}_{\alpha},{N}_{i\alpha}$
, thus
$${G}_{\alpha}={\oplus}_{n\in \omega}R{x}_{\alpha n}^{0},{N}_{i\alpha}={\oplus}_{n\in \omega}R{x}_{\alpha n}^{i}\text{for}i=1,2.$$
If
$\alpha <\beta \in {\omega}_{1},n\in \omega $
, then
$${x}_{\alpha n}^{i}={\sum}_{m\in \omega}{a}_{\beta nm}^{i\alpha}{x}_{\beta m}^{i}\text{for some}{a}_{\beta nm}^{i\alpha}\in R.$$
Since
${z}_{i}\in {G}_{0}$
, also
${z}_{i}={\sum}_{m\in \omega}{b}_{m}^{i}{x}_{0m}^{0}$
for some
${b}_{m}^{i}\in R$
. Now we can choose a subring
${R}^{\prime}$
of
$R$
of cardinality
${\aleph}_{1}$
containing all these coefficients
${a}_{\beta nm}^{i\alpha},{b}_{m}^{i}$
.
Let
$${G}_{\alpha}^{\prime}={\oplus}_{n\in \omega}{R}^{\prime}{x}_{\alpha n}^{0},{N}_{i\alpha}^{\prime}={\oplus}_{n\in \omega}{R}^{\prime}{x}_{\alpha n}^{i}\text{for}i=1,2$$
be the corresponding free
${R}^{\prime}$
module. We have
${\aleph}_{1}$
filtrations
${G}^{\prime}={\cup}_{\alpha \in {\omega}_{1}}{G}_{\alpha}^{\prime}$
and
${N}_{i}^{\prime}={\cup}_{\alpha \in {\omega}_{1}}{N}_{i\alpha}^{\prime}$
; this uses that
${R}^{\prime}$
is sufficiently saturated! Moreover, passing from
$G$
to
${G}^{\prime}$
it is easy to see that also
${G}^{\prime}$
is a strongly
${\aleph}_{1}$
free
${R}^{\prime}$
module. Also
${z}_{i}\in {G}_{0}^{\prime}$
and
${\phi}_{i}^{\prime}={\phi}_{i}\upharpoonright {G}^{\prime}$
is an
${R}^{\prime}$
homomorphism
${G}^{\prime}\u27f6Q$
.
By another back and forth argument, using that
${\phi}_{i}:G\u27f6Q$
are epimorphisms (so enlarging
${R}^{\prime}$
) we may assume that
${\phi}_{i}^{\prime}$
maps
${G}^{\prime}$
onto the quotient field
${Q}^{\prime}$
of
${R}^{\prime}$
. By Observation 3.1 we also assume that
${R}^{\prime}$
is a principal ideal domain. Thus
${G}^{\prime}/{N}_{i}^{\prime}\sim ={Q}^{\prime}$
for
$i=1,2$
, and we can apply Theorem 1.3
$\left(ii\right)$
. The
${R}^{\prime}$
module
${G}^{\prime}$
is a Crawley
${R}^{\prime}$
module and there is
${\Theta}^{\prime}\in Au{t}_{{R}^{\prime}}{G}^{\prime}$
with
${N}_{1}^{\prime}{\Theta}^{\prime}={N}_{2}^{\prime}$
.
Finally we extend
${\Theta}^{\prime}$
(uniquely) to
$\Theta \in Au{t}_{R}G$
such that
${N}_{1}\Theta ={N}_{2}$
:
There is a cub
$C\subseteq {\omega}_{1}$
such that
${\Theta}_{\alpha}^{\prime}={\Theta}^{\prime}\upharpoonright {G}_{\alpha}^{\prime}\in Au{t}_{{R}^{\prime}}{G}_{\alpha}^{\prime}$
for all
$\alpha \in C$
; moreover
${\Theta}^{\prime}={\cup}_{\alpha \in C}{\Theta}_{\alpha}^{\prime}$
and
${G}^{\prime}={\cup}_{\alpha \in C}{G}_{\alpha}^{\prime}$
. These
${R}^{\prime}$
automorphisms
${\Theta}_{\alpha}^{\prime}$
act on the free
${R}^{\prime}$
module
${G}_{\alpha}^{\prime}={\oplus}_{n\in \omega}{R}^{\prime}{x}_{\alpha n}^{0}$
; they extend naturally to
$R$
automorphisms
${\Theta}_{\alpha}$
of the free
$R$
module
${G}_{\alpha}={\oplus}_{n\in \omega}R{x}_{\alpha n}^{0}$
by tensoring with
$R$
. Clearly
$\Theta ={\cup}_{\alpha \in C}{\Theta}_{\alpha}$
and
${N}_{1}\Theta ={N}_{2}$
.
(This also shows that
$G$
is
${\aleph}_{1}$
presented, i.e. it can be expressed as the quotient of two free
$R$
modules of rank
${\aleph}_{1}$
.) If
$R$
is a principal ideal domain with quotient field
$Q$
countably generated over
$R$
, then there is a multiplicatively closed subset
$\mathbb{S}$
of
$R\backslash \left\{0\right\}$
such that
${\mathbb{S}}^{1}R=Q$
. We may assume that
$1\in \mathbb{S}$
and choose an enumeration
$\{{s}_{i}:i\in \omega \}=\mathbb{S}$
with
${s}_{0}=1$
.
Moreover
${q}_{n}={\prod}_{i\le n}{s}_{i}$
. In particular
$0={\cap}_{i\in \omega}R{q}_{i}$
, so
$R$
is an
$\mathbb{S}$
ring; see [
7]
. Examples are the principal ideal domains of
$p$
adic integers
${J}_{p}$
with
$\mathbb{S}=\{{p}^{n}:n\in \omega \}$
and counter examples are polynomial rings
$\mathbb{Z}\left[X\right]$
over
$\mathbb{Z}$
in sets X of uncountably many, commuting variables. If
$R$
is such an
$\mathbb{S}$
ring, then we can construct Griffith's strongly
${\aleph}_{1}$
free
$R$
module of rank
${\aleph}_{1}$
: Let
$F={\oplus}_{i\in {\omega}_{1}}R{e}_{i}$
be the free
$R$
module of rank
${\aleph}_{1}$
and
$\widehat{F}$
its
$\mathbb{S}$
adic completion; note that the
$\mathbb{S}$
topology on
$F$
is Hausdorff by the above. Let
$E\subseteq {\omega}_{1}$
be a stationary subset of limit ordinals and choose for any limit ordinal
$\delta \in {\omega}_{1}$
a ladder, i.e. a strictly increasing sequence of successor ordinals
${\delta}_{n}(n\in \omega )$
with limit
$\delta $
. We consider the elements
$${v}_{\delta}^{k}={\sum}_{n\ge k}\left({q}_{n}{q}_{k}^{1}\right){e}_{{\delta}_{n}}\in \widehat{F}$$
and let
$$G=\langle F,R{v}_{\delta}^{k}:\delta \in E,k\in \omega \rangle $$
which is a pure submodule of
$\widehat{F}\cap {\prod}_{i\in {\omega}_{1}}R{e}_{i}$
. It is easy to check, that
$G$
is strongly
${\aleph}_{1}$
free with
$\Gamma $
invariant
$\Gamma \left(E\right)\ne 0$
, hence not free. Thus the next corollary follows from Theorem 3.2 .
Corollary 3.3
(ZFC + MA +
${\aleph}_{1}<{2}^{{\aleph}_{0}}$
) If
$R$
is a principal ideal domain with quotient field
$Q\ne R$
countably generated over
$R$
, then there are
${\aleph}_{1}$
free but not free Crawley
$R$
modules of rank
${\aleph}_{1}$
.
Corollary 1.4 follow immediately from Corollary 3.3 and Theorem 3.2 .
References

A. L. S. Corner and R. Göbel, Prescribing endomorphism algebras, a unified treatment, Proc. London Math. Soc. 50 (1985), 447–479.

A. L. S. Corner, R. Göbel and B. Goldsmith, The Existence of torsion–free, nonfree Crawley Groups is Undecidable, to be submitted (2004).

K. J. Devlin, S. Shelah, A weak version of
$\u25ca$
which follows from
${2}^{{\aleph}_{0}}<{2}^{{\aleph}_{1}}$
, Israel J. Math. 29 (1978) 239 – 247.

P. Eklof and A. Mekler, Almost Free Modules, Settheoretic Methods, NorthHolland, 1990.

L. Fuchs, Abelian Groups, Publishing House of the Hungarian Academy of Science, Budapest 1958.

L. Fuchs, Infinite Abelian Groups, Vol. I and II, Academic Press, 1970 and 1973.

R. Göbel and J. Trlifaj, Approximations and endomorphism algebras of modules , Walter de Gruyter, Berlin 2004.

T. Jech, Set Theory, Academic Press 1978.

C. Megibben, Crawleys problem on the unique
${\omega}_{1}$
elongations of
$p$
groups is undecidable, Pacific J. Math. 107 (1983) 205 – 212.

A. H. Mekler, S. Shelah,
$\omega $
elongations and Crawley's problem, Pacific J. Math. 121 (1986) 121 – 132.

A. H. Mekler, S. Shelah, The solution of Crawley's problem, Pacific J. Math. 121 (1986) 133 – 134.

S. Shelah, Infinite abelian groups, Whitehead problem and some constructions, Israel J. Math. 18 (1974) 243 – 256.

S. Shelah, Proper and Improper Forcing, Springer Verlag, Berlin 1997.