Injective pairs in perfect rings (Q1268671): Difference between revisions

A pair \((eA,Af)\) of a right ideal \(eA\) and a left ideal \(Af\) in a left perfect ring \(A\) (with unity) is called an \(i\)-pair if \(e,f\) are local idempotents, \(eA_A\) and \(_AAf\) have essential socles such that \(\text{soc}(eA_A)\cong fA/fJ(A)\), \(\text{soc}({_AAf})\cong Ae/J(A)e\), where \(J(A)\) is the Jacobson radical of \(A\). The authors prove that for an \(i\)-pair \((eA,Af)\) the following conditions are equivalent: (i) \(_{eAe}(eA)\) is Artinian; (ii) \((Af)_{fAf}\) is Artinian; (iii) both \(eA_A\) and \(_AAf\) are injective modules. It is proved also that (i) \(eA_A\) is \(A\)-simple-injective (i.e., for any \(N_A\leq A_A\) every \(\theta\colon N_A\to eA_A\) with \(\text{Im }\theta\) simple can be extended to some \(\phi\colon A_A\to eA_A\)) if and only if there exists a local idempotent \(f\in A\) such that \((eA,Af)\) is an \(i\)-pair and \(r_{Af}(l_{eA}(M))=M\) for any submodule \(M\) of \((Af)_{fAf}\); (ii) \(eA_A\) is injective if it is \(A\)-simple injective and has finite Loewy length.