If we are reasoning with rules that are guidelines rather than absolute statements of truth, we need a mechanism for selecting the most plausible conclusion. There may be a choice of rules that can be applied in some situations, and using one may result in a conclusion that is contradictory to that obtained using another.

This paper presents a mechanism where there is a reliability factor associated with each rule. The monotonic property means that the reliability factor given to any conclusion will never be more than the reliability factor of any intermediate step used to reach it. The mechanism is used to derive a set of expressions, each of which has a final reliability factor.

Although this is presented as a general framework, the focus of the paper is on using it for ontology matching. Suppose we have two vocabularies that describe the same situation. Ontology matching means determining for each term in one what the equivalent is in the other. So using the mechanism described in this paper results in a set of pairs of matching terms, with a reliability factor for each match.

The paper gives data showing its use on some established benchmarks for ontology matching, with it performing well compared to other systems used for ontology matching.