Tue, 7 October 2025, 1:25 pm MT
A strong Maltsev class of varieties is a class of varieties for which there exists a finitely presented variety (i.e. a variety in a finite signature satisfying finitely many identities) whose interpretability class is a lower bound to every variety in the interpretability order, while a Maltsev class of varieties is a class for which there exists a countably infinite descending chain of finitely presented varieties and each variety in the Maltsev class is bounded below by some of the chain. A Maltsev term is the original example of a strong Maltsev condition and the identities that define it define the class of congruence permutable varieties. On the other hand, the classes of congruence distributive and congruence modular varieties are not strong Maltsev classes, but rather Maltsev classes (the terms for distributivity were discovered by Jónsson and the terms for modularity were discovered by Day).
There has been for some years a fertile area of research which connects Universal Algebra to the complexity theory of constraint satisfaction. While many of these results are largely important for complexity theory, some of them have been important for algebra. In particular, Siggers discovered that the class of locally finite Taylor varieties is captured by a finite package of identities. In 2015, Olsak made the surprising discovery that in fact the class of all Taylor algebras is strong Maltsev. A natural next question was: what about the class of congruence meet semidistributive varieties? This is an important class in fixed template finite domain CSP, since the class of locally finite congruence meet semidistributive varieties is exactly the class of varieties which correspond to CSP templates which can be solved with local consistency methods.
Actually, Kozik, Krokhin, Valeriote, and Willard proved that locally finite congruence meet semidistributive varieties are characterized by a finite package of identities. We show that, unlike the class of Taylor varieties, this finiteness does not lift to the general case, that is, there does not exist a strong Maltsev condition that captures every congruence meet semidistributive variety.