Institute of Mathematics

Abstract:  Discrete partially ordered sets can be turned into distance spaces in several ways. The distance functions may or may not satisfy the triangle inequality and restrictions of the distance to finite chains may or may not coincide with the natural, difference-of-height distance measured in a chain. It is shown that for semilattices the semimodularity ensures the good behaviour of the distances considered. The Jordan-Dedekind chain condition, which is weaker than semimodularity, is equivalent to the basic criterion that the graph-theoretic distance (realized by zig-zagging up and down freely in the poset to connect two points) is compatible with distances measured on chains by the relative height. Semimodularity is shown to be equivalent to the validity of the triangle inequality of a restricted graph-theoretic distance, called the up-down distance. The fact that the up-down distance corresponds to the computation of degrees of kinship in family trees leads to the observation that the less familiar canon-law method of computation corresponds also to a mathematically well behaved Chebyshev-type distance on discrete semilattices. For the Chebyshev distance also semimodularity is shown to imply the validity of the triangle inequality. The reverse implication fails, but assuming the validity of the triangle inequality, the semimodularity is shown to have a local characterization by a forbidden six-element subsemilattice. Like in the classical case of real spaces, the Chebyshev semilattice distance is shown to be the limit of a converging sequence of distances, all of them verifying the triangle inequality if the semilattice is semimodular.