Re: SUO: Re: Re: Conformance (actually, "standards profiles")
Hi all,
I am travelling for some time, and it is not so easy to
follow your discussions, but I did.
And I can't rest without answering to the ongoing discussion
with Seth.
See below
Josiane
>
> From: "West, Matthew R SITI-ITPSIE"
<Matthew.R.West@IS.shell.com>
>
> > > Yes I think think standards profiles would work and
might be
> > > quite eligant.
> > > Are these profiles chosen by the user, or prescribed
by the
> > > standards body?
> > > I prefer the former. But I really should recuse
myself from further
> > > discussion of standards as I don't even believe in
creating a standard
> > > ontology. I just want a dictionary in which to look
up terms
> > > such that
> > > others will understand what I am trying to say.
> >
> > MW: But that is what a standard is ... something that is
used by a lot
> > of people.
>
> Ok, I agree. Then the meaning of conformance is that you
are speaking using
> only the dictionary definition of the words. But then can
you make up your
> own words? ... because those made up words would not be in
the dictionary
> and hence not in conformance?
>
> Seth Russell
>
It seems to me that the main problem underlying these
discussions raises the question of 'meaning schema' and
'contextual dimensions'.
I have not my data with me, but I had already given you at
least one exemple (in the course of a discussion with Adam
if I remember well).
'Meaning schema' and 'contextual dimensions' are both
procedural invariants. While procedural instructions
composing the 'meaning schema' act upon concepts, procedural
instructions composing every contextual dimension act as an
interface between concepts and physical world. Moreover the
'meaning schema' is the invariant through 'contextual
dimensions', and each contextual dimensions define a class
of equivalent context.
But in order to draw out more clearer semantic concepts
corresponding to these notions, one returns always to the
same point, viz to choice a way for coding protocols.
Two sorts of coding are necessary : one at level of objects
and situation (with SUMO if I understood well), the other at
syntatic level (the common part to KIFF and Chris's Menzel
system, as our previous discussions showed it).
Then two lattices would be setted one in order to represent
the physical and situational level of every occurrences of a
word, and the other lattice to represent the syntactic
co-text of the word.
(In my previous researches I was using correspondance
analysis)
So these two independant lattices could be matched together
and they can give a formal characterization of the
contextual dimensions and latter of the meaning schema.
Please let you send your answer to my usual address
josiane.caron@mshs.univ-poitiers.fr
Josiane