Laws in a class of groupoids

From MaRDI portal





In an earlier paper [Bull. Aust. Math. Soc. 26, 385-392 (1982; Zbl 0492.08008)] the first author introduced a class of groupoids derived from commutative semigroups. The previous paper showed that such groupoids have no finite basis for their laws; this paper investigates the laws that hold in all groupoids of the class. Let \({\mathcal S}=(S,+)\) be a commutative groupoid, and \(\alpha,\;\beta\) two commuting endomorphisms of \(\mathcal S\). A new operation \(\mu\) is defined on \(S\) by \(xy\mu=x\alpha+x\beta\) and it is the groupoid \({\mathcal G}(\alpha,\beta)=(S,\mu)\) (or \(\mathcal G\) for short) that is the object of attention. Since it is only the laws that hold in all such groupoids that are of interest here, \(\mathcal S\) can be taken to be the free commutative semigroup of countable rank, and \(\alpha,\;\beta\) to be freely commuting. An example of such a law (in fact, the shortest nontrivial law) is the medial law \(xy\mu zt\mu^ 2=xz\mu yt\mu^ 2\). An infinite sequence of laws is defined as follows. Let \[ u_ m=x_ 1x_ 2\dots x_ m\mu^{m-1}, v_ m=x_{m+1}x_{m+2}\dots x_{2m-1}\mu^{m- 2}x_{2m}\mu,\;w_ n=u_ mv_ m\mu,\;\hbox{ where } n=2m. \] Let \(w_ n'\) be the word obtained from \(w_ n\) by interchanging the variables \(x_{m-1}\) and \(x_{2m-2}\); then the main result of the paper is: Theorem. Each of the laws \(w_ n=w_ n'\) for \(n=2m\geq 6\) holds in \({\mathcal G}\) and is independent of all shorter laws of \(\mathcal G\). An immediate consequence of this is that the laws of \(\mathcal G\) do not have a finite basis.











This page was built for publication: Laws in a class of groupoids

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1182933)