Re: SUO: Re: SUO Ballot with 2 Questions
Hello, Pierre.
I don't see what 'absolutism' or 'relativism' have to do with a SUO, as in
your remarks below, although I suppose they might have. Unless you were
thinking of the Temporal Logic disputes - 3D vs. 4D.
But very well, let us say that a SUO is not, after all, a Standard Upper
Ontology, but something less comprehensive, as you indicate -- a Standard
Upper Scientific Ontology. (Or is it to be a Standard Upper Engineering
Ontology?)
What sort of science shall we include? Mathematical physics, common-sense
physics? Biology? Chemistry? Quantum physics? Economics -- is it a science?
Psychology -- is it a science?
What are the essential common vocabularies of these sciences? What are the
boundaries of science? Do you know a formal means of indicating them? I
don't, and I don't believe it exists, but I see great difficulties in trying
to settle such a question, and I would have to defer to specialists in every
branch. While it's well known that experts often disagree -- and experts in
logic disagree about the nature of logic.
One might attempt to formalize some sciences in set theory (a la Suppes,
e.g.) but I don't know what special vocabulary one could use for all of the
candidate sciences beyond that particularly specialized language. Using set
theory only, though, would leave us with a 'simple' ontology of sets (or
classes?), so that the SUO work, in one sense, might be quickly done. But
even if such a limited SUO were done in this way, what would be done about
discovering and resolving disagreements among competing axiomatizations?
What would have been accomplished, lacking that?
On the other hand, to introduce vocabulary 'directly' is to invite endless
possibilities of inconsistency (or irresolution), as Cyc, for instance, does
with CycL. Logically speaking, how is the introduction of specialized
vocabulary to remain 'neutral' on ontological questions? By a very carefully
crafted hierarchy of universal hypotheticals? Is that possible or practical,
or particularly useful? In any event, at some point not far off,
axiomatizations competing in actual use will have to face existential
counterclaims.
I think that the way forward is (in part) to create a mechanism for
registering and recording the ontological possibilities, and for resolving
differences as they might be discoverable. That, to me, seems to me to be
the best course toward 'accepting our limitations', which you also
recommend, and of acknowledging the very practical and actual realities of
context, viewpoint, and of process and change in our states of knowledge.
I don't object to anyone trying to create a SUO, which is why I voted for
the SUMO proposal, but I don't expect it to be interestingly useful very
soon to anyone in the 'real world'.
I am sure that there are many varieties of interesting but incompatible
SUOs, which is why I voted for the second proposal.
Best,
Jay
----- Original Message -----
From: "Pierre Grenon" <pierre.grenon@ifomis.uni-leipzig.de>
To: "Jay Halcomb" <jhalcomb8@attbi.com>
Cc: <standard-upper-ontology@ieee.org>
Sent: Tuesday, June 10, 2003 10:16
Subject: Re: SUO: Re: SUO Ballot with 2 Questions
> Hi,
>
> This sounds like a weird argument to me. I thought we were doing science
and
> engineering maybe... The theological questions raised below are not of
this
> order at all and the analogy doesn't hold imho.
>
> There is no claim that aiming at an ontology (whatever its structure!)
rather
> than a series of alternative is an attempt to derive _the_ ultimate and
> absolute ontology. But it seems reasonable to me to aim at something which
is
> the closest to the best of our knowledge - (we is a relevant community,
> scientific one in particular).
>
> To use your example for another purpose, an SUO would not tell you whether
> there are gods and how many of them if any. It could tell you that there
are
> agents and that these agents have beliefs if you want. This seems both
true and
> relevant to the SUO. From there you can build the ontology of whichever
> mythology you like or whichever belief system consistent with one side of
the
> debate. Gods are agents, people have are agents, agents have beliefs,
people
> may believe in gods and so on... An SUO ought to be neutral on narrow
domain
> issues.
>
> I think we have both a problem of scope and method with the SUO. SUO is
general
> (enough to trace over a number of the most contentious distinction) and it
> should reflect whatever the state of our knowledge is conventional, maybe
among
> certain communities of expert (although this is what tends to motivate
splits).
>
>
> The trouble is not whether we will have a definitive list of all
conceivable or
> possible entities. This is not only irrelevant to the SUO, it is harmful.
The
> trouble is that we are putting in place some ad hoc machinery which
reflect a
> doctrinal position and some obscur obligation to acknowledge the will of a
> couple of systems. Relativist poseurs on this list are wimps: if they
really
> believed in what they profess they should just accept the limitations of
our
> knowledge and live with it. Accept that an ontology like science and
knowledge
> may change and hopefully progress. A standard ought to be amendable by
> principle and this is precisely how an SUO could progress, rather than by
> apologizing for its existence.
>
> To speak as some of my dearest AI visionaries, if an SUO is really
something
> which must reflect a purpose and some contingency, it seems to me that if
we
> come up with one which serves its purpose for a couple of decades, we
should be
> more than happy. If it outlives us, people should start worrying about it.
Why
> would somebody who believes that truth may not be absolute would do
something
> as foolish as to presume the nature of what people in more than 20 years
from
> now will find to be a good SUO? And why bother them with a shitload of
> alternative which may turn out to be irrelevant in 5 years from now? This
> reflect some of the incoherence of the kind of relativism advocated in
this
> list. I'm afraid there is more work toward achieving glory than toward
> achieving any goal stated in the PAR.
>
> Best
> Pierre
>
>