Model-theoretic aspects of \(\Sigma\)-cotorsion modules (Q876378)

A module \(M\) is said to be cotorsion if \(\text{Ext}^1(M,F)=0\) for every flat module \(F\). Pure-injective modules are cotorsion and this paper is part of a programme of the authors to extend the well-developed theory of pure-injectives to cotorsion modules. Say that a module \(M\) is \(\Sigma\)-cotorsion if for every \(\kappa\) the sum, \(M^{(\kappa)}\), of \(\kappa\) many copies of \(M\) is cotorsion. The analogous condition of \(\Sigma\)-pure-injectivity is equivalent to the descending chain condition on pp-definable subgroups (also called finite matrix subgroups) so it is reasonable to ask whether being \(\Sigma\)-cotorsion is equivalent to the dcc on certain kinds of pp-definable subgroups. The authors show that for countable rings this is so and they identify the particular kinds of pp-definable subgroups as those given by divisibility conditions. Their results are more general but beyond the context of countable rings the situation is more complicated. Another result which specialises nicely to countable rings is that, for such rings, it is enough to check the defining condition with \(\kappa =\aleph_0\) -- something which is true without restriction for \(\Sigma\)-pure-injectivity.