Entities usually have several features, so it’s often possible to categorise the same entities into two or more different hierarchies. Facet theory provides a way of handling this cleanly and systematically.
If we categorise the four animals below using the facet of habitat, the duck is closer to the frog than to the hen, because frogs and ducks are both amphibious.
However, on the facet of zoological classification, the the duck is closest to the hen, since both are birds, and the duck is also more closely related to the stag than to the frog.
A common mistake when creating classification systems is to try to fit two or more facets within a single hierarchy. It’s much cleaner and simpler to use a separate hierarchy for each facet.
Facet theory is widely used by libraries for categorising information sources. The same concept is used within graph theory and within knowledge representation, in concepts such as an is-a hierarchy versus a part-of hierarchy.
This is highly relevant to laddering, where people often switch between facets without realising it, leading to potential confusion.