Inference does not properly handle subtypes, especially for multiple parent types

[franzlst]

In our custom OWL ontology, we have defined the following type hierarchy (excerpt):

owl:Thing
	Entity
		Physical
			Object
				Agent
					Robot
						LbrArm
							Lbr3
							Lbr4
				SelfConnectedObject
					CorpuscularObject
						Artifact
							Device
								Robot
									LbrArm
										Lbr3
										Lbr4

Here Robot shows up twice in the hierarchy, as it is defined as subtype of both Agent and Device:

<rdf:Description rdf:about="http://purl.org/ieee1872-owl/cora-bare#Robot">
    <rdfs:subClassOf rdf:resource="http://purl.org/ieee1872-owl/sumo-cora#Agent"/>
    <rdfs:subClassOf rdf:resource="http://purl.org/ieee1872-owl/sumo-cora#Device"/>
</rdf:Description>

Now I want to get all information about Lbr3:

SELECT  ?s ?p ?o
WHERE {  <https://ontology.SHORTENED.de/SHORTENED#Lbr3> ?p ?o  }

On my system, I get the following result:

s	p	o
	rdf:type	rdfs:Class
	rdf:type	owl:Class
	rdfs:subClassOf	http://purl.org/ieee1872-owl/cora-bare#Robot
	rdfs:subClassOf	https://ontology.SHORTENED.de/SHORTENED#LbrArm
	rdfs:subClassOf	https://ontology.SHORTENED.de/SHORTENED#Lbr3
	rdfs:subClassOf	rdfs:Resource

I would have expected that the result includes also subClassOf entries for owl:Thing, Entity, Physical, Object, Agent and Device, but the entries stop at Robot.

Related to this issue is when querying information about an instance of Lbr3, named example_lbr:

SELECT  ?s ?p ?o
WHERE {  <https://ontology.SHORTENED.de/SHORTENED#example_lbr> ?p ?o  }

s	p	o
	rdf:type	http://purl.org/ieee1872-owl/cora-bare#Robot
	rdf:type	owl:NamedIndividual
	rdf:type	https://ontology.SHORTENED.de/SHORTENED#LbrArm
	rdf:type	https://ontology.SHORTENED.de/SHORTENED#Lbr3
	rdf:type	t744
	rdf:type	t752

The result shows two unnamed nodes (t744 and t752). I could imagine that these relate to Device and Agent. But what don't they have names and why are the types not listed up to Thing?

[franzlst]

Interestingly, a colleague of mine has the very same setup (including the *.properties file). For the same query

SELECT  ?s ?p ?o
WHERE {  <https://ontology.SHORTENED.de/SHORTENED#example_lbr> ?p ?o  }

his result shows the full type hierarchy:

p	o
rdf:type	http://purl.org/ieee1872-owl/sumo-cora#Artifact
rdf:type	http://purl.org/ieee1872-owl/sumo-cora#CorpuscularObject
rdf:type	http://purl.org/ieee1872-owl/sumo-cora#Entity
rdf:type	http://purl.org/ieee1872-owl/sumo-cora#Object
rdf:type	http://purl.org/ieee1872-owl/sumo-cora#Physical
rdf:type	http://purl.org/ieee1872-owl/sumo-cora#SelfConnectedObject
rdf:type	http://purl.org/ieee1872-owl/cora-bare#Robot
rdf:type	owl:NamedIndividual
rdf:type	https://ontology.SHORTENED.de/SHORTENED#LbrArm
rdf:type	https://ontology.SHORTENED.de/SHORTENED#Lbr3
rdf:type	t1053
rdf:type	t1122

So in summary, there are three issues here:

• on my setup, the inference regarding type hierarchy is not complete
• the inference for the same (what else could be different?) setup is inconsistent
• unions do not show up as named nodes