The Grothendieck conditon \(AB5^*\) and generalizations of module distributivity (Q1390412)

For a certain cardinality \(\omega\geq 2\) a lattice \(L\) is said to be \(\omega\)-distributive, if each of its nonempty subsets of a cardinality not exceeding \(\omega\) has a greatest lower bound and \(y\cup(\bigcap_{\lambda\in\Lambda}y_\lambda)=\bigcap_{\lambda\in\Lambda}(y\cup(\bigcap_{\mu\in\Lambda\setminus\{\lambda\}} y_\mu))\) for all elements \(y\in L\) and families \((y_\lambda)_{\lambda\in\Lambda}\) with \(|\Lambda|=\omega\). A module \(M_R\) is called \(\omega\)-distributive if the lattice of its submodules is \(\omega\)-distributive. A subfactor \(S\) of a module \(M_R\) is a submodule of a factor module of \(M\) (i.e. \(S=A/B\subseteq M/B\)). The notion of independent family \((y_\lambda)_{\lambda\in\Lambda}\) of submodules of \(X_R\) is dualized in different forms (which coincide if \(X_R\) is an \(AB5^*\)-module). The main result (Theorem 1) describes the property of \(\omega\)-distributivity of \(M\) by independent and co-independent families of submodules of subfactors of \(M\) (14 equivalent conditions). The countably distributive modules (\(\omega=\aleph_0\)) are characterized by simple modules (Theorem 2), in particular \(M\) is countably distributive iff for any simple module \(P_R\) and every submodule \(K\subseteq M\), there exists a finitely generated submodule \(T\subseteq K\) such that \(\Hom_R(K/T,P)=0\). The ring \(R\) is right purely semisimple iff every right \(R\)-module is a direct sum of \(AB5^*\)-modules (or linearly compact modules, or \(\aleph_0\)-distributive modules).