ontology-summit
[Top] [All Lists]

Re: [ontology-summit] OntologyFrameworkDraftStatementfortheOntologySummi

To: Ontology Summit 2007 Forum <ontology-summit@xxxxxxxxxxxxxxxx>
From: "Deborah L. McGuinness" <dlm@xxxxxxxxxxxxxxxx>
Date: Sun, 22 Apr 2007 09:45:41 -0700
Message-id: <462B9135.90804@xxxxxxxxxxxxxxxx>
i have tried to stay out of this since there already is a lot of 
discussion on this point  ... but i also agree that i have always looked 
at the word ontology with multiple senses.
i was doing some homework on existing ontology definitions and noticed 
the listing on dictionary.com
http://dictionary.reference.com/browse/ontology    (01)

there is precedent with mentioning a computer science sense of ontology  
but the one i find interesting is the on from the free online dictionary 
that states it is from philosophy - "An explicit formal specification of 
how to represent the objects, concepts and other entities that are 
assumed to exist in some area of interest and the relationships that 
hold among them.
"    (02)

it goes on to expand that  but the simple sentence is one i have used.    (03)

(the expansion they give gives enough for us to talk about for the next 
month or year  so all i was really pointing out was the single sentence 
but for completeness i include their expansion:  "For AI 
<http://dictionary.reference.com/browse/AI> systems, what "exists" is 
that which can be represented. When the knowledge 
<http://dictionary.reference.com/browse/knowledge> about a domain 
<http://dictionary.reference.com/browse/domain> is represented in a 
declarative language 
<http://dictionary.reference.com/browse/declarative%20language>, the set 
of objects that can be represented is called the universe of discourse 
<http://dictionary.reference.com/browse/universe%20of%20discourse>. We 
can describe the ontology of a program by defining a set of 
representational terms. Definitions associate the names of entities in 
the universe of discourse 
<http://dictionary.reference.com/browse/universe%20of%20discourse> (e.g. 
classes, relations, functions or other objects) with human-readable text 
describing what the names mean, and formal axioms 
<http://dictionary.reference.com/browse/axioms> that constrain the 
interpretation and well-formed use of these terms. Formally, an ontology 
is the statement of a logical theory 
<http://dictionary.reference.com/browse/logical%20theory>.
A set of agents <http://dictionary.reference.com/browse/agents> that 
share the same ontology will be able to communicate about a domain of 
discourse without necessarily operating on a globally shared theory. We 
say that an agent commits to an ontology if its observable actions are 
consistent with the definitions in the ontology. The idea of ontological 
commitment is based on the Knowledge-Level 
<http://dictionary.reference.com/browse/Knowledge-Level> perspective.
"    (04)


John F. Sowa wrote:
> Mike, Bill, Chris, Steve, Leo, and Barry,
>
> MU> These two word senses [from the M-W dictionary] pretty
>  > much do cover the difference between what I was calling
>  > philosophical ontology, vs. IT/CS ontologies.
>
> That should not be surprising, because every IT/CS ontology
> depends on some prior *philosophical* analysis -- unfortunately,
> it's sometimes rather bad philosophy.
>
> BA> The Merriam-Webster definition is, IMHO, pretty good, except
>  > for the bit about "specifically abstract entities" and the focus
>  > on language.  The latter is more forgivable since it is, after all,
>  > systems grounded in more-or-less formal language that we're talking
>  > about.  The former seems confused - why the focus on "specifically
>  > the kinds of abstract entities" while admitting (presumably)
>  > non-abstract entities.
>
> I agree that the word 'specifically' is confusing.  It was actually
> written "specif.", which is a common abbreviation in that dictionary
> for a more specific sense that follows.  That phrase could be changed
> to "including abstract entities".
>
> Note that the word 'language' was used in the phrase "language systems".
> That definition from 1971 was written by a philosopher (M-W does use
> editors who are experts in the subject matter) who was well aware of
> the work on formal languages in the first-half of the 20th century.
> To clarify that point, we could add the phrase "formal and informal"
> at the end.
>
> With those two revisions, definition 2 becomes:
>
>   2. a theory concerning the kinds of entities, including abstract
>      entities, to be admitted to a language system, formal or
>      informal.
>
> The language system, for example, could be Common Logic and all its
> dialects.  The entities "admitted" to that system would be everything
> in the domain of quantification.  The theory would be all the axioms
> that that refer to those entities.
>
> CP> Even when it is about describing a situation -- it is not always
>  > clear how reference works. David Armstrong gives as an example
>  > the statement that "there are at least two people in the room"
>  > -- when there are a lot more. What does the statement refer to
>  > (e.g. which two people?) -- you have to go through quite a few
>  > contortions to rescue reference.
>
> Those "contortions" are handled very precisely by model theory,
> If anybody asks "Which two do you mean?"  The answer is simple:
> "Any two -- your choice."
>
> CP> So what seems to me to characterise a model of an ontology
>  > is a desire to map the "things in the world" directly via
>  > reference – and that language, concepts, etc do not necessarily
>  > share that desire.
>  >
>  > I am not sure that this desire has been made explicit in the
>  > current Ontology Framework Draft Statement for the Ontology
>  > Summit -- and I think it might usefully do so.
>
> I agree that a few words would be useful, and I suggest some
> words in my response to Leo (at the end of this note).
>
> SN> As for me, I doubt that there's anything invariant about
>  > the soup, and I suspect that whatever may appear to be
>  > invariant cannot be relied upon to remain so.
>
> I was using the word 'invariant' in the sense of mathematics,
> physics, and computer science:  a relationship (described by
> some mathematical or logical expression) that remains unchanged
> under some transformation.
>
> In physics, for example, there can be constant, even chaotic,
> motion, but the focus of the subject is on what remains invariant
> under various transformations.  Examples include things like mass,
> energy, momentum, angular momentum, etc.
>
> When we're talking about knowledge soup, the invariants would be
> patterns that remain constant under various kinds of translations
> from one language to another.  (And by the way, different invariants
> may be associated with different kinds of transformations.)
>
> LO> "Theories", they think they understand because they've heard
>  > the word as referring to scientific theories, but they don't
>  > really know what a theory is.
>  >
>  > So I start off using "concept" and tell them simultaneously
>  > that it is a placeholder for the thing in the world, etc.
>  > Then I build up to theories, in fact logical theories.
>
> I think you can say something short and understandable without
> raising dubious or at least debatable issues about concepts, etc.
> At the end of this note is my suggestion.
>
> BS> Am I right in thinking that you want a 3-level theory, here,
>  > with concepts serving as intermediaries between terms and
>  > entities?  If so, why is this intermediary level necessary?
>  > How does it help?  How, in particular, does it help pedagogically,
>  > given that (demonstrably) people find the term 'concept' so
>  > difficult to understand?
>
> I agree that we should not raise any of those issues in the
> summary.  I just checked the M-W and Longman's dictionaries for
> a definition of 'concept'.  M-W gave a long list of options,
> and Longman's didn't attempt to define the word.  Following is
> their entry:
>
>     "a general idea, thought, or understanding."
>
> The lexicographers who wrote that definition had no desire to
> enter the tar pit.
>
> Following is a suggested definition of the two senses of the
> word 'ontology' and two sentences of explanation.  The best
> way to clarify that definition is to give examples.
>
> John
> ________________________________________________________________
>
> [In the following definition, the first sense is taken from
> the _Longman Dictionary of Contemporary English_ and the second
> from M-W, as modified above in response to Bill Andersen.]
>
> The word 'ontology' is used in two senses:
>
>   1. The branch of philosophy concerned with the nature of
>      existence and the relations between things.
>
>   2. A theory concerning the kinds of entities, including
>      abstract entities, to be admitted to a language system,
>      formal or informal.
>
> In computer systems, the language can be any version of logic,
> such as Common Logic, RDF, OWL, or many others.  A theory is
> a collection of statements in some version of logic that is
> used to characterize the entities and relations of some domain.
>  
> _________________________________________________________________
> Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/ 
> Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
> Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
> Community Files: http://ontolog.cim3.net/file/work/OntologySummit2007/
> Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
> Community Portal: http://ontolog.cim3.net/
>       (05)


_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/ 
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2007/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
Community Portal: http://ontolog.cim3.net/    (06)
<Prev in Thread] Current Thread [Next in Thread>