Invariant means on Boolean inverse monoids (Q284646)

Let \(S\) be an inverse semigroup with zero. \(S\) is called distributive if it has binary joins of compatible pairs of elements, and multiplication distribute over such joins. If in addition the semilattice \(E(S)\) of idempotents is a Boolean algebra then \(S\) is called Boolean. The Boolean inverse semigroups and inverse monoids are considered. It is said that \(S\) has an invariant mean if there is a function \(\mu:E(S)\to [0,\infty)\) satisfying the following conditions: 1) for any \(s\in S\), \(\mu (s^{-1}s)=\mu (ss^{-1})\); 2) if \(e\) and \(f\) are orthogonal idempotents (i.e., \(ef=0\)), then \(\mu (e\vee f)=\mu (e)+\mu (f)\). The authors declare that the fundamental question that interests them in their paper ``is the existence or non-existence of invariant means on Boolean monoids''. The main result is a characterization of when a Boolean monoid admits an invariant mean. The obtained conditions generalize the classical Tarski alternative. Some examples are given.