Mutually essentially pseudo-injective modules. (Q263018)

Generalizing the notion of quasi-injective module, pseudo-injective modules and essentially pseudo-injective (e.p.injective in short in this review) modules have been studied already. An \(R\)-module \(M\) is said to be e.p.\(N\)-injective (for an \(R\)-module \(N\)) if every monomorphism from an essential submodule \(A\) of \(N\) into \(M\) can be extended to a homomorphism from \(N\) into \(M\). An \(R\)-module \(M\) is said to be e.p.injective if \(M\) is e.p.\(M\)-injective. Two modules \(M,N\) are said to be mutually e.p.injective if each is e.p.injective relative to the other. The authors, in this paper, study mutually e.p.injective modules and investigate their properties. They prove that an \(R\)-module \(M\) is e.p.\(N\)-injective if and only if \(f(N)\subseteq M\) for each monomorphism \(f\colon E(N)\to E(M)\). This generalizes the similar result for e.p.injective modules. In Theorem 2.5 they prove that every submodule of an \(R\)-module \(M\) is e.p.injective if and only if every essential submodule of \(M\) is e.p.injective. Here in proving \((3)\Rightarrow(1)\) the authors use the fact that every direct summand of an e.p.injective module is e.p.injective which is Corollary 2.8. The authors should have proved this result as a lemma prior to Theorem 2.5. In Theorem 2.6(3), \(M,N\) are ``mutually'' e.p.injective should be assumed. In Theorem 2.11 the authors prove that a ring \(R\) is QF if and only if the class of all projective \(R\)-modules and strongly e.p.injective \(R\)-modules is socle fine, that is, any two \(R\)-modules in this family are isomorphic if and only if their socles are isomorphic. Next they prove that a ring \(R\) is semisimple if and only if the class of all strongly e.p.injective \(R\)-modules is socle fine. Here in the proof of \((3)\Rightarrow(1)\), \(\mathrm{Soc}(R_R)\) should be strongly e.p.injective which is not obvious. In the proof of the Theorem 2.15 the authors use the result that for an \(R\)-module \(M\), if \(M\oplus E(M)\) is e.p.injective then \(M\) is injective. They should have mentioned this as a consequence of Theorem 2.7(2). In this theorem, in (1), the latter conclusion should be modified as ``Furthermore, in either of the cases, \(R\) is a right Noetherian, right V-ring''. The authors use the result that if every semisimple \(R\)-module is injective then \(R\) is a right Noetherian, right V-ring. The V-ring part is well known but Noetherian ring part is not so well known. For this Vamos's theorem should have been quoted. After studying properties of e.p.injective modules, the authors study relationship of right e.p.injectivity of a ring \(R\) with that of a triangular matrix ring for an \((S,R)\)-bimodule \(M\), corner ring \(eRe\), for an idempotent \(e\) in \(R\) and polynomial ring \(R[X]\).