Work at SourceForge, help us to make it a better place! We have an immediate need for a Support Technician in our San Francisco or Denver office.

Close

#22 RelationshipType extent and OneToOneFunction

v0.7
open
nobody
KB Content (9)
5
2003-02-20
2003-02-20
No

Relation has 265 subcolllections, RelationshipType has
262 instances. I suggest that all subcollections of
Relation should be instances of RelationshipType.
Which are missing?

Mt : UniversalVocabularyMt
EL Query :
(and
(genls ?X Relation)
(not
(isa ?X RelationshipType)))

Answer ?X
*[Explain #3] ModalOperator
*[Explain #2] NLMorphologyPredicate
*[Explain #1] ActionPredicate
*[Explain #0] ComplexActionPredicate

Note that negation by failure was enabled and inference
in InferencePSC did not find a result.

If we add these four collections to the extent of
RelationshipType the number of instances would be 266.
Consequently, there must be one instance of
RelationshipType, which is not a subcollection of
Relation contrary to the monotonic assertion in
UniversalVocabularyMt:
(typeGenls RelationshipType Relation).
The one is OneToOneFunction:

Mt : UniversalVocabularyMt
EL Query :
(and
(not
(genls ?X Relation))
(isa ?X RelationshipType))

Answer ?X
*[Explain] OneToOneFunction

Discussion

  • Casey McGinnis
    Casey McGinnis
    2005-02-10

    Logged In: YES
    user_id=1085367

    Assuming that RelationshipType (to be renamed as
    RelationType, incidentally) means the same as (SpecsFn
    Relation), you are right. The problem seems to be fixed
    now, however. (At least in OpenCyc and full Cyc; I haven't
    checked ResearchCyc.) Moreover, our policy is to avoid
    reifiying "powercollections" (analogous to powersets), i.e.
    the collection of all subcollections of some collection. We
    can always denote the powercollection of a collection using
    SpecsFn (which is unreifiable). There are still numerous
    powercollections in the KB, but most if not all of them
    should eventually be eliminated.