On dual pairs and simple-injective modules (Q1974170): Difference between revisions

Let \(P\) and \(Q\) be rings, \(M\) a left \(P\)-module and \(N\) a right \(Q\)-module. If there is a \(P\)-\(Q\)-bilinear map \(\varphi\colon M\times N\to V\) for some \(P\)-\(Q\)-bimodule \(V\), \((M,N)\) is said to be a pair (with respect to \(\varphi\)) and is called a dual pair if \(lr(K)=K\) and \(rl(L)=L\) hold for any submodule \(K\) of \(M\) and any submodule \(L\) of \(N\), where \(r\) and \(l\) mean right annihilator (with respect to \(\varphi\)) in \(N\) and left annihilator (with respect to \(\varphi\)) in \(M\), respectively. A right \(Q\)-module \(N\) is called quasi-dual if there is a left \(P\)-module \(M\) such that \((M,N)\) is a dual pair and a ring \(R\) is called right quasi-dual if the right \(R\)-module \(R\) is quasi-dual. In this paper, it is shown that, in the first section, any linearly compact module is quasi-dual, any quasi-dual module over a semiperfect ring has finite Goldie dimension and if \(R\) is a right quasi-dual ring, then \(R\) is semiperfect and any finitely generated right \(R\)-module has finite Goldie dimension. In the second section, using the notion of simple-injectivity, the authors improve previous results and characterize dual rings.