Re: SUO: Universal Time, other universals, and cultural contextof SUO.

["John F. Sowa" <sowa@bestweb.net>]

Pat,

Every formal ontology, including your work on temporal axioms,
starts with notions abstracted from NL.  The only difference between
the ones used in expert systems and the ones used for NLP is in the
problems they are designed to solve.

In other words, there is no sharp break between Kn. Rep. and
NL semantics.  It is a continuum, with the differences primarily
determined by the purpose to which the representation is being
applied.

See my other note about Iris T's work (and the pointer to her
web page in my corrected version, with the spelling Tommelein).

A few comments:

> I greatly respect Miller and I am sure that WordNet is a valuable
> resource for someone, but it is utterly and entirely useless for
> anything to do with ontology. (Do you recall the discussions at
> Heidelberg with Piek about EuroNet, and how  useless they were? None
> of the linguistically inspired categories make ontological sense.)

Yes, indeed.  They need to learn more about logic (although some of
them, such as Martin van den Berg do, but he didn't read the right
logic books).

> >And there are quite a few logicians who hang out with computational
> >linguists, especially when working on theories of modality,
> >tense logics, etc.
> 
> Right. If one is interested in the logical structure of NL - and I
> have every respect for those who are, let me say, I find it
> fascinating myself - then this is the thing to do. But that is one
> game, and ontology/KR is another game.

If done right, those games begin to merge.  That is the main theme
of paper I am writing, which I mentioned in my reply to Nicola.
That is one of the major reasons why I think that Nicola isn't
getting anywhere with his approach -- he is too hung up on an
obsolete, late 20th century version of philosophy (a version that
was made obsolete by the work that the 20th century philosophers
never read).

John