SUO: RE: Latest Version of Merged Ontology
Hi Robert,
Here's an attempt to spell out the procedure I used in
merging/aligning the ontologies that make up the "Merged Ontology". The
first step was to convert all of the ontologies listed on the SUO Candidate
Content web page (http://ltsc.ieee.org/suo/refs.html) into SUO-KIF. Once
this was accomplished, I selected the most general ontology (John Sowa's
upper level ontology) as the base ontology. I then mapped concepts/axioms
from the remaining ontologies into the base ontology. In this mapping
process, there were two possible cases, each in turn with two subcases.
1. New content (Nothing in the merged ontology corresponds with the
concept/axiom to be mapped)
a. Useful/Relatively Noncontroversial content. In this subcase, the
new concept/axiom is deemed to be useful and not to violate any cherished
philosophical principles. Once I decided that the new concept/axiom should
be included, it was simply a matter of finding a place for it in the
ontology. In some cases, this required creating some intermediate
levels of concepts between existing concepts and new content. These
intermediate levels were identified in an earlier email directed at
John Sowa on "trouble spots" in the merged ontology. An example of this
subcase is the subhierarchy of numbers that was taken from the KIF-Numbers
ontology on the Ontolingua server.
(subclass-of RealNumber Number)
(subclass-of RationalNumber RealNumber)
(subclass-of PositiveRealNumber RealNumber)
(subclass-of NegativeRealNumber RealNumber)
(subclass-of Integer RationalNumber)
(subclass-of EvenInteger Integer)
(subclass-of OddInteger Integer)
(subclass-of NaturalNumber Integer)
(subclass-of NonnegativeInteger Integer)
(subclass-of NegativeInteger Integer)
(subclass-of PositiveInteger Integer)
(subclass-of PositiveNumber Number)
(subclass-of NegativeNumber Number)
(subclass-of ComplexNumber Number)
b. Useless/Controversial content. In this subcase, the new
concept/axiom is judged (by me) to be either not useful or too
controversial for incorporation into a schema that we hope will have both
broad application and wide acceptance. This is of course a very subjective
judgement, but ontology merging/alignment is probably much more of
an art than a science. An example of content that was thought to be too
controversial for inclusion is the following hierarchy of normative notions
developed by myself. It is too controversial, because there are many ways
of categorizing such notions, and some have suggested that a version of
Deontic Logic would be more appropriate for capturing the required
connections.
;; (subclass-of Normative Entity)
;; (subclass-of NormativeProposition Normative)
;; (subclass-of NormativeProposition Proposition)
;; (subclass-of Obligation NormativeProposition)
;; (subclass-of Agreement NormativeProposition)
;; (subclass-of JudgementOfEtiquette NormativeProposition)
;; (subclass-of AestheticJudgement NormativeProposition)
;; (subclass-of InstitutionalObligation Obligation)
;; (subclass-of PersonalObligation Obligation)
;; (subclass-of ReligiousObligation InstitutionalObligation)
;; (subclass-of LegalObligation InstitutionalObligation)
2. Old Content (There is some degree of overlap between the merged ontology
and the concept/axiom to be mapped)
a. Perfect Overlap. In this case, the terms may differ but the new
concept has the same semantic content as a concept already in the merged
ontology or, with respect to axioms, there is a logical equivalence
between the new axiom and an existing axiom. An example of this sort of
case occurs in the various mereotopological theories. Some philosophers
use 'part-of' as the primitive notion to frame their axioms, some use
'overlaps', and still others use some notion of 'connection'. However, in
many cases, axioms across these different systems are logically equivalent.
b. Partial Overlap. In this case, there is some overlap in meaning
between the new content and existing concepts or axioms in the merged
ontology. An example of this is the case role hierarchy. Although
John Sowa's system of case roles was part of the base ontology (and, hence,
the original merged ontology), I tried to simplify it somewhat to make it
more like other systems of case roles. Both systems follow:
John Sowa's Subhierarchy under 'Role':
;;(subclass-of PrehendingEntity Role)
;;(subclass-of PrehendedEntity Role)
;;(subclass-of Composite PrehendingEntity)
;;(subclass-of Correlative PrehendingEntity)
;;(subclass-of Correlative PrehendedEntity)
;;(subclass-of Component PrehendedEntity)
;;(subclass-of Whole Composite)
;;(subclass-of Substrate Composite)
;;(subclass-of Part Component)
;;(subclass-of Property Component)
;;(subclass-of Piece Part)
;;(subclass-of Participant Part)
;;(subclass-of Stage Part)
;;(subclass-of Attribute Property)
;;(subclass-of Manner Property)
;;(subclass-of Determinant Participant)
;;(subclass-of Source Participant)
;;(subclass-of Product Participant)
;;(subclass-of Immanent Participant)
;;(subclass-of Initiator Determinant)
;;(subclass-of Initiator Source)
;;(subclass-of Resource Source)
;;(subclass-of Resource Immanent)
;;(subclass-of Goal Determinant)
;;(subclass-of Goal Product)
;;(subclass-of Essence Product)
;;(subclass-of Essence Immanent)
;;(subclass-of Agent Initiator)
;;(subclass-of Recipient Goal)
;;(subclass-of Beneficiary Recipient)
;;(subclass-of Completion Goal)
;;(subclass-of Destination Goal)
;;(subclass-of Duration Resource)
;;(subclass-of Effector Initiator)
;;(subclass-of Experiencer Goal)
;;(subclass-of Instrument Resource)
;;(subclass-of Location Essence)
;;(subclass-of Matter Resource)
;;(subclass-of Medium Resource)
;;(subclass-of Origin Initiator)
;;(subclass-of Path Resource)
;;(subclass-of Patient Essence)
;;(subclass-of PointInTime Essence)
;;(subclass-of Result Goal)
;;(subclass-of Start Initiator)
;;(subclass-of Theme Essence)
The Revised Case Roles (the version that appears in the latest
release of the merged ontology):
(subclass-of Predicate Role)
(subclass-of ComponentRelation Predicate)
(instance-of part-of ComponentRelation)
(subclass-of component-of part-of)
(subclass-of external-covering component-of)
(subclass-of internal-component-of component-of)
(instance-of property-of ComponentRelation)
(subclass-of BinaryRelation Predicate)
(subclass-of TernaryRelation Predicate)
(subclass-of VariableArityRelation Predicate)
(subclass-of CaseRole BinaryRelation)
(subclass-of IntentionalRelation BinaryRelation)
(subclass-of IntentionalRelation IntentionalEntity)
(subclass-of PropositionalAttitude IntentionalRelation)
(subclass-of ObjectAttitude IntentionalRelation)
(instance-of stage-of ComponentRelation)
(subclass-of attribute-of property-of)
(subclass-of manner-of property-of)
(instance-of source-of CaseRole)
(instance-of agent CaseRole)
(instance-of recipient CaseRole)
(instance-of beneficiary CaseRole)
(instance-of completion CaseRole)
(instance-of destination CaseRole)
(instance-of duration CaseRole)
(instance-of effector-of CaseRole)
(instance-of experiencer CaseRole)
(instance-of instrument CaseRole)
(instance-of located-at CaseRole)
(instance-of matter CaseRole)
(instance-of medium CaseRole)
(instance-of origin CaseRole)
(instance-of path CaseRole)
(instance-of patient CaseRole)
(instance-of exists-at CaseRole)
(instance-of result CaseRole)
(instance-of theme CaseRole)
I hope this note helps to clarify the process I used in merging/aligning
ontologies. Let me know if you want additional details.
-Ian
> -----Original Message-----
> From: Robert E. Kent [mailto:rekent@ontologos.org]
> Sent: Wednesday, January 17, 2001 9:15 AM
> To: Ian Niles
> Cc: SUO
> Subject: Re: Latest Version of Merged Ontology
>
>
> Ian,
>
> Since I am an advocate for a building blocks approach to ontology
> construction, and understand that the mortar between the
> blocks must be
> strong and resilient in order to adequately support the
> building, I would
> like to request that we spell out very clearly how the
> component ontologies
>
> -- in this case Sowa's upper ontology, Russell and Norvig's
> ontology, Casati
> and Varzi's theory of holes, Allen's temporal axioms, the relatively
> noncontroversial elements of Smith's and Guarino's respective
> mereotopologies, the KIF formalization of the Core Plan
> Representation, a
> Simple-Time ontology, a Standard-Units ontology, an Agents
> ontology, the
> Numbers ontology (from the Ontolingua server), and the
> Natural-Kinds and
> Positions ontology (authored by the Ontology Group at
> ITBM-CRN), but in
> general any ontologies --
>
> are put together. You mention the two operations or steps of
> (1) merging and
> (2) alignment. In order that everyone comprehends what is
> involved here, can
> we give these an explicit and detailed explanation with illustrious
> examples? And are there other hidden and/or useful operations
> for building a
> composite ontology?
>
> Robert E. Kent
> rekent@ontologos.org
>
>
> ----- Original Message -----
> From: "Ian Niles" <iniles@teknowledge.com>
> To: "Standard-Upper-Ontology (E-mail)"
> <standard-upper-ontology@ieee.org>
> Sent: Friday, January 12, 2001 10:38 AM
> Subject: SUO: Latest Version of Merged Ontology
>
>
> > Hi All,
> >
> > It's attached. This version incorporates innumerable minor
> > corrections, but the biggest differences between it and the
> version I sent
> > out three weeks ago are as follows:
> >
> > 1. Incorporates six new ontologies, viz. Simple-Time,
> Standard-Units,
> > Agents, and Numbers (from the Ontolingua server), and
> Natural-Kinds and
> > Positions (authored by the Ontology Group at ITBM-CRN).
> The original
> forms
> > of all of these ontologies can be accessed from the SUO
> Candidate Content
> > page (http://ltsc.ieee.org/suo/refs.html).
> >
> > 2. The temporal axioms have been heavily revised on the
> basis of Pat
> Hayes'
> > comments.
> >
> > 3. The subhierarchy of high-level concepts (borrowed from
> John Sowa) has
> > been simplified in an attempt to eliminate concepts that
> are of purely
> > philosophical interest. This is a work in progress, and it
> was prompted
> by
> > both Pat Hayes and Chris Menzel.
> >
> > 4. The case role portion of the ontology (also borrowed
> from John Sowa)
> has
> > also been simplified.
> >
> > I'd like to extend a special thanks to Pat Hayes and Chris Menzel
> > for their detailed, helpful, and challenging remarks on the previous
> version
> > of the merged ontology, and I'd like to encourage everyone
> to comment on
> the
> > new version. Let me know if anyone has any difficulty viewing the
> attached
> > file.
> >
> > -Ian
> >
> >
> >
> >
>
>