Re: SUO: RE: Link Grammar and Parser - RRGs
SUO: "Suggested Upper Ontology".
I guess the "upper" is something that people find hard to stick with.
> Part of what is therefore necessary is that we begin to share
> ideas and our understanding of our different domains, as well as
> begin to construct a common understanding and lexicon.
This cannot be correct. This means, paraphrased, that this group wants to
talk about all domains and then model them. This is Cyc. If SUO
starts making concrete suggestions for any particular domain then
it is off-course *unless* this is being done by experts in that field.
One should know whether one is an expert in a field or not, and
refrain from proposing ontology modules in areas where one is not.
One can be a very interested amateur in some area and follow
the discussion--but this cannot seriously be being proposed
as a basis for ontology modules of such areas.
> there'd be little point in everyone "sticking to their
> areas of expertise", as that'd be like watching The McLaughlin
> Group, where each pundit spouts their dogma and nobody listens
> to the others.
I have to disagree and repeat my reference to the U of SUO.
Taking my own area, "language", as an example, any discussion there has
to fit
*both* into the prerequisites established on whatever formal ontological
grounds have been agreed upon in a group activity such as SUO (and I
believe that there at least there *is* some real expertise in the group)
*and* into the expertise available in the field. If one is not an expert
in that field
then one should not take too much bandwidth discussing the
details of the module or hypothesising other strange augmentations
or theoretical positions. It appears to me that this kind of discussion
is pursued far too often in the SUO list and detracts from the
neat stuff.
> another equally important
> part is to build a community for such an effort.
Discussions in areas that are clearly not particularly represented in
terms of expertise are not a good way of constructing a community
that will be in a position to deliver. Those who are will turn off.
Perhaps the situation is
worse in the language area than it is in other areas, and so my
negative comments should not be generalized to tar all the
discussions with the same brush: it is very common to hear
opinions that suggest that because everyone can speak everyone
is automatically a "linguistic" expert. This kind of error is not
so likely to occur when dealing with obviously more technical
areas. If I started sounding off on an ontology of business
processes, I trust that I would be put right quick and would have
the sense to keep quiet and go and learn more before taking
up more time on the SUO list with my own education.
If an activity of the SUO list is raising every member's awareness
and expertise in all of the areas under discussion, then it is
certainly a worthy enterprise: but I have no idea what that has to do with
getting an upper ontology with some well constructed "not quite
so upper" modules in place. The theory of quantum physics does
not rest on a group of non-experts having agreed about its
"understanding and lexicon" for that theory.
I would be interested to know if the notion of a "cooperative
pedagogy" is really going to structure SUO activities from now
on rather than having expert working groups that go out on
hunting parties.
> I'm sorry if you feel you are above this kind
> of activity
above? below? With respect to upper ontology, language descriptions
are "below". With respect to pedagogy, I do that when I teach
linguistics and computational linguistics (ie. most of the time);
I come to the SUO list to do
ontology or see ontology being done: *not* to teach linguistics 101.
My idea of the list and the group may well be far too narrow:
or perhaps the list needs to split to cover the particular
kinds of discourses that it wants to support. Only some of
these discourses can be considered as leading up to
SUO content in terms of building an upper ontology
or lattice of such ontologies.
Perhaps these could be clearly earmarked
or signposted so that one does not have to worry that the
latest suggestion is going to sweep the floor and become
a SUO-standard in some moment of enthusiasm
(e.g., hey guys, lets designate Relational
Grammar as the way to structure linguistic information....).
Separating out the activities and their intended *function* within
the activities as a whole could also give cooperative
pedagogy an appropriate home in the whole and one
that would indeed be positive in the way you suggest.
Knowing the function of a discourse contribution is
a vital prerequisite for communication of any kind: apparently
the function "building a SUO" is one that many
of the contributions to the list are not directly
concerned with. It would be good to have these
explicitly designated as such.
> It's been fashionable lately to criticize the activities
well, maybe it's not only due to "fashion"....
John B.