Varieties of Boolean inverse semigroups

From MaRDI portal



Abstract: In an earlier work, the author observed that Boolean inverse semi-groups, with semigroup homomorphisms preserving finite orthogonal joins, form a congruence-permutable variety of algebras, called biases. We give a full description of varieties of biases in terms of varieties of groups: (1) Every free bias is residually finite. In particular, the word problem for free biases is decidable. (2) Every proper variety of biases contains a largest finite symmetric inverse semigroup, and it is generated by its members that are generalized rook matrices over groups with zero. (3) There is an order-preserving, one-to-one correspondence between proper varieties of biases and certain finite sequences of varieties of groups, descending in a strong sense defined in terms of wreath products by finite symmetric groups.


Recall, that an inverse semigroup is a semigroup \(S\) where every \(x\in S\) has a unique inverse, that is, an element \(x^{-1}\) such that \(x = xx^{-1}x\) and \(x^{-1} = x^{-1}xx^{-1}\). It is well known that the class of all inverse semigroups as algebras of the type \(\{\cdot, ^{-1} \}\) is a variety. A generalized Boolean algebra \(G\) is a relatively complemented distributive lattice with largest element \(1\) (that is, an element \(a\in G\) has a complement in any interval \([x,1]\) that contains \(a\)). Elements \(x\) and \(y\) in an inverse semigroup with zero are orthogonal if \(x^{-1}y = xy^{-1} = 0\). An inverse semigroup \(S\) with zero is Boolean if the set of all idempotent elements of \(S\) is a generalized Boolean algebra and any two orthogonal elements \(x\) and \(y\) in \(S\) have a join with respect to the natural ordering, then denoted by \(x \oplus y\). The class of Boolean inverse semigroups is not defined as a variety of algebras of the signature \(\{\cdot, ^{-1}, 0, \oplus \}\) since the orthogonal join \(\oplus\) is only a partial operation. The author introduced in [Refinement monoids, equidecomposability types, and Boolean inverse semigroups. Cham: Springer (2017; Zbl 1447.20005)] two full operations the skew difference \(\backslash\) and the skew addition \(\nabla\) and proved in that the structures \((S; \cdot, ^{-1}, 0, \backslash ,\nabla)\) can then be axiomatized by a finite number of identities, whose models are called \textit{biases}. The main aim of the paper is describing varieties of biases in terms of varieties of groups. The author proves that: ``(1) Every free bias is residually finite. In particular, the word problem for free biases is decidable. (2) Every variety of biases is generated by its fully group-matricial members and every proper variety of biases contains a largest finite symmetric inverse semigroup, and it is generated by its members that are monoids of generalized rook matrices over groups with zero. (3) There is an order-preserving, one-to-one correspondence between proper varieties of biases and certain finite sequences of varieties of groups, descending in a strong sense defined in terms of wreath products by finite symmetric groups.



Cites work









This page was built for publication: Varieties of Boolean inverse semigroups

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1663513)