A criterion for a finite endoprimal algebra to be endodualisable (Q5950765)

In the paper the strategy for finding endoprimal algebras due to B. A. Davey and J. G. Pitkethly is further explored in the finite case. An algebra \({\mathbf M}=(M,F)\) is called \(k\)-endoprimal \((k\geq 1)\) if every \(k\)-ary End\textbf{M}-preserving function on \textbf{M} is a term function on \textbf{M}. Algebras which are \(k\)-endoprimal for every \(k\geq 1\) are called endoprimal. A finite algebra \textbf{M} is said to be endodualisable if End\textbf{M} yields duality on the quasivariety \({\mathbb I}{\mathbb S}{\mathbb P}\)(\textbf{M}). The paper is built on the work of \textit{B. A. Davey} and \textit{J. G. Pitkethly} [Algebra Univers. 38, 266-288 (1997; Zbl 0934.08004)] and further elucidates the concept of endoprimality and its relationship to endodualisability. The Retraction Test Algebra Lemma is used as a tool to show that, in many quasivarieties, endoprimality is equivalent to endodualisability for finite algebras which are suitably related to finitely generated free algebras.