I'm trying to define different states for object properties between OWL classes in my ontology. What already works is creating a new object property every time I'm having a new state for it.
So this would be the way im solving it:
Concept --- hasConcept --> Concept
Now I want to see all the archived Concepts of a Concept. So whenever I'm archiving one Concept I'm changing the object property from hasConcept
to the following new object property:
Concept --- hasArchivedConcept --> Concept
The problem with this solution is that I'll have to create the same hasArchived*
object property for many other classes in my ontology.
Concept --- hasTerm --> Term
Concept --- hasArchivedTerm --> Term
and so on.
In order to deny that effort of having to create multiple object properties every time I need a new state, I'm looking for another smooth semantic OWL way to fullfill that requirement.
OWL isn't the best representation for this due to being limited to binary predicates. You're either trying to represent something is true at a known time, or in a hypothetical or known situation, which in AI is known as a Fluent (see Situation Calculus):
colour(car, red, time1). % time1 can be substituted for a situation reference
colour(car, blue, time2).
So we'd need ternary relations. The solution from Situation Calculus, is to reify the fluent:
holds(colour(car, red), situation0).
Each predicate is now only arity 2, but OWL doesn't allow a triple as a subject.
The simplest solution for OWL is from the recommendation for representing n-ary relations in OWL, use-case 1:
colour_at_time(car, _ct1).
colour(_ct1, red).
time(_ct1, t1).
There's a disadvantage to such a work-around: you are sacrificing reasoning and clarity in queries.
There also exists an Ontology based on Situation Theory plus several event based ones with regards to time. Good places for research are: ODPs, Semantic Sensor Network Ontology because of observations, Prov-O because changes to documents over time, and BFO v1 for qualities, processes and the now removed object properties.