Dans un courrier dat=E9 du 07/06/99 18:21:50 , charles a =E9crit :
> Now, if the loglang well-defines approximately 10 words,
> it can do propositional logic; 10 more, and it can do
> predicate calculus. If nouns are defined carefully with
> respect to set theory, ontologies should not be a problem;
they are not anymore.
> WordNet and many other projects have done so. Animals are
> not really all that mysterious ... Adjectives and verbs
> are more problematically interesting, but at least
> several classes of words can be well-defined.
> =20
you can define them provided you STOP equating action-and-state with "verb"=20
and substance with "noun". try "predicate" and "substantive" instead and=20
ponder about what "argument" is. do, be, have. you're next to enlightment=20
then. and consider romance linguists' works regarding semantics. read=20
yoshiko's EL page. it's naive but very instructive as to what non-europeans=20
and europeans like me have in their heads. read the EL's "verb" section=20
especially. if you "do" not understand it then it "is" a pity and i "have"=20
problems. ;-)
> Though the mathematician Couturat had a major role in
> designing parts of Ido, and Peano (famous for his set theory)
> invented an auxlang (Interlingua, the first of that name)
the best ial for latin peoples. oops ! i withdraw that. oh no ! it's too lat=
e=20
: i've sent it already ;-)
> neither was especially logical in its design or constructions.
> =20
did you ever wonder why ?
mathias=20