Hereditary radicals and quasiradicals in universal algebras (Q2266035): Difference between revisions

For the variety of unitary right \(S\)-modules over a commutative ring with identity \textit{J.-M. Maranda} [Trans. Am. Math. Soc. 110, 98--135 (1964; Zbl 0121.26601)] established a bijective correspondence between the radicals resp. quasiradicals (in other terminology: preradicals) which are strongly hereditary and certain sets of right ideals of \(S\). The basic idea of the present paper is to generalize Maranda's results considering \(S\) as the one-generator free right \(S\)-module. Because of the lack of a good correspondence between subalgebras and congruences in general, two ways of generalization arise: 1) (quasi)radicals \(R(A)\in \mathrm{Con}\, A\) strongly hereditary in the sense that for every subalgebra \(B\) of \(A\) the congruence \(R(B)\) is the restriction of \(R(A)\) to \(B\); 2) 0-(quasi)radicals \(r(A)\in \mathrm{Sub}\, A\) strongly hereditary in the sense of \(r(B)=r(A)\cap B\) for every \(B\in \mathrm{Sub}\, A\) (for the second way, the existence of a nullary operation ''0'' which is a subalgebra is required). Strong heredity is equivalent in case 2) to the property ''\(a\in r(A)\) iff \(a\in r(<a>)''\), in case 1) to ''(a,b)\(\in R(A)\) iff \((a,b)\in R(<a,b>)''\) (where \(<M>\) denotes the subalgebra generated by \(M\) in \(A\)). Considering the occurring one- resp. two generator subalgebras as homomorphic images of free algebras, the author obtains: in case 1) a bijective correspondence between the strongly hereditary (quasi)radicals on the considered variety \({\mathcal V}\) and certain sets of congruences of the two-generator free algebra in \({\mathcal V}\); characterizations of the semisimple and the radical classes are given. For case 2) he obtains a bijective correspondence between the strongly hereditary 0-quasiradicals and some sets of congruences of the one-generator free algebra in \({\mathcal V}\). Under some restrictions concerning the congruences ''generated'' by subalgebras, analogues of this result are obtained for strongly hereditary 0-radicals resp. \(0^*\)-(quasi)radicals (for \(0^*\) \(r(A)\) is supposed to be a full congruence class in \(A\)).