An easy test for congruence modularity (Q451080): Difference between revisions

Let \(\Sigma\) be a set of identities. \(\Sigma\) is inconsistent if \(\Sigma \models x \approx y\), otherwise \(\Sigma\) is consistent. \(\Sigma\) is idempotent if for every function symbol \(F\) appearing in \(\Sigma\), it is the case that \(\Sigma \models F(x, x,..., x) \approx x\). A term \(t\) is linear if it has at most one occurrence of a function symbol. An identity \(s \approx t\) is linear if both \(s\) and \(t\) are linear. There are two well-known theorems, \textit{A. Day}'s theorem [Can. Math. Bull. 12, 167--173 (1969; Zbl 0181.02302)] and \textit{H. P. Gumm}'s theorem [Arch. Math. 36, 569--576 (1981; Zbl 0465.08005)], which each characterize congruence modularity. The authors of the paper under review describe an easy way to determine whether the realization of a set of idempotent identities guarantees congruence modularity or the satisfaction of a non-trivial congruence identity. Namely, for a set \(\Sigma\) of idempotent identities, they define the notion of a derivative, \(\Sigma '\), which is a superset of idempotent identities in the same language. One of their main results is that \(\Sigma\) axiomatizes a congruence-modular variety if the derivative of \(\Sigma\) is inconsistent. Another main result is that \(\Sigma\) axiomatizes a variety that satisfies some nontrivial congruence identity if its \(n\)-th derivative is inconsistent for some \(n\). The two final results show that for a set \(\Sigma\) of idempotent linear identities, the derivative test is a necessary and sufficient condition to determine if \(\Sigma\) defines a variety that is congruence-modular or satisfies a nontrivial congruence identity.