The paper studies a mathematical structure called a pseudocomplemented lattice, which is an ordered set where every element has a unique "best complement" called its pseudocomplement. These structures have well-understood and elegant properties. The challenge is that many naturally occurring ordered sets do not have this nice pseudocomplementation property, so the authors ask: can we find a useful substitute that works more broadly?
The authors introduce an operator that assigns to each element not a single best complement, but rather a whole set of maximal elements that are "orthogonal" to it, meaning they share no common part above zero. They call this the p-operator. They show that as long as the ordered set satisfies a condition called the Ascending Chain Condition, which roughly means you cannot have infinitely long strictly increasing sequences of elements, this operator is well-defined and behaves in ways that closely mirror true pseudocomplementation. They work out the detailed properties of this operator specifically for semilattices and lattices, which are two commonly studied types of ordered structures.
Finally, the authors use this operator to define an equivalence relation on the ordered set, grouping together elements that look the same from the perspective of the p-operator. This idea was inspired by classical work of the mathematician Glivenko on distributive lattices. They then investigate when this grouping is compatible with the lattice structure in a strong enough sense to define a well-behaved quotient structure, and they identify conditions under which that quotient has especially nice properties such as being complemented or modular.