Neutrabelian algebras (Q2226985)

Any cogruences \(\alpha\) and \(\beta\) of an algebra in a congruence modular variety satisfy the condition: \(0\leq [\alpha,\beta]\leq\alpha\wedge\beta\). An algebra is called \textit{abelian}, if \([\alpha,\beta]=0\) and is called \textit{neutral} if \([\alpha,\beta]=\alpha\wedge\beta\) [\textit{R. Freese} and \textit{R. McKenzie}, Commutator theory for congruence modular varieties. Cambridge: Cambridge University Press. London: London Mathematical Society (1987; Zbl 0636.08001)]. The authors of the paper introduce the notion of \textit{neutrabelian} algebras in a congruence modular variety. According to their definition, an algebra is \textit{neutrabelian SI} if it is subdirectly irreducible; the centralizer \(\nu\) of the monolith is comparable to all other congruences; for congruences \(\alpha,\beta\leq \nu\), the commutator \([\alpha,\beta]=0\); and \([\alpha,\beta]=\alpha\wedge\beta\), otherwise. An algebra is \textit{neutrabelian} if every its subdirectly irreducible quotient is a neutrabelian SI. The authors show that a finite algebra in a congruence modular variety is neutrabelian if and only if it has centalizers split at \(0\). As a consequence of this fact they obtain that each finite algebra with cube term and every subalgebra neutrabelian is dualizable. Moreover, a finite algebra in a finitely decidable congruence modular variety is dualizable.