[Top] [All Lists]

Re: [ontac-forum] Problems of ontology -- and terminology

To: ONTAC-WG General Discussion <ontac-forum@xxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sun, 21 May 2006 11:26:13 -0400
Message-id: <44708695.3050009@xxxxxxxxxxx>
Dear Matthew,    (01)

I realize that you and most of the other people in this
group are well aware of the differences between an
ontology and a terminology.    (02)

But many people have been claiming to "align ontologies"
by using WordNet.  That implies the only things common
to the categories that have been aligned are the WordNet
synsets -- unless, and this is a very big *UNLESS* --
they give a formal proof that all inferences in the two
systems are logically equivalent.  Have you ever seen
anyone give such a proof for a large system?  I haven't.    (03)


MW> This is a problem for people who come from a linguistic
 > background, but as an engineer, I deal with reality, not
 > how people describe it.    (05)

But I would add that everybody on earth comes from a "linguistic
background" -- namely, their own native language, which for
better or worse always affects how they think about everything.    (06)

In any case, it's meaningless to talk about anybody's claims
about reality that are independent of descriptions or depictions.
Whatever we know about reality is only known through our senses
and our theories -- both of which involve some kind of iconic
or symbolic representation.    (07)

JS>> A lot of people have aligned their terminologies, but nobody
 >> has ever aligned any large ontology with any other in the
 >> sense that all inferences with one are *identical* to all
 >> the inferences with the other.    (08)

MW> You do not align legacy systems. You map between them. The
 > purpose is to ensure that what is mapped has the same meaning
 > in each system.    (09)

This leads to another fundamental principle:    (010)


As soon as the ink is dry on any design, it has become a legacy
system.  As soon as version 1.01 is released, version 1.0 is a
legacy.  But even within any large organization, people have begun
work on version 2.0 long before they ever release version 1.0.    (012)

Furthermore, nobody has ever given a formal proof that the terms
that describe any version of a large system have "the same meaning"
in one version and the next.  And if that is true of the "same"
system, you can be absolutely certain that it is true of any two
systems that have been or will be developed independently.    (013)

All the talk about how things will be better when everybody has
adopted some particular version of an upper ontology is just
"pie in the sky".  And following is the complete line from
the song by Joe Hill:    (014)

    Pie in the sky when you die -- it's a lie.    (015)

Mapping between independently developed systems and legacy systems
has always been and will always be a *permanent* reality.  There
will never be a case when everybody is using identical large-scale
ontologies.  The most that can ever be claimed is that two systems
agree on a narrowly circumscribed task or set of tasks.    (016)

Any other claim is pure pie in the sky -- in simple terms, a lie.    (017)

John    (018)

Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki: 
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG    (019)
<Prev in Thread] Current Thread [Next in Thread>