SUO: *Dated Material 19 Jan 2002
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤
Item 1. Quod Vitae
| In November 1619, I had a dream involving
| the Seventh Ode of Ausonius, which begins
| 'Quod vitae sectabor iter'
| [What road in life shall I follow].
|
| René Descartes, "Experimenta", CSM 1, page 4.
When I think about where we are going with this SUO business,
I imagine a time ten years in the future, when I envision that
we might have some sort of "ontology understanding system" (OUS),
on the Web or Whatever, and when we ask it a question of the form
"What is ... ?" it gives us a sensible and a genuinely informative
answer, simple in the beginning, but one that would lead us to more
complete answers if we so desire to pursue them.
When I think back to the variety of experiences that I have had working
with different kinds of logical systems, and especially since the time
when I first thought that it would be a good idea to work toward the
development of a "mathematically oriented theorem automater" (MOTA),
or a "mathematically oriented theorem organizer" (MOTO), then it
occurs to me that some of my experiences must have been rather
different from some of the other people in the group, mainly
because I long ago abandoned as utterly hopeless many of
the ways that I see them spending their time and energy
pursuing, and though I sympathize more than they know,
there does not seem to be anything I can do about it,
since they just go on in that way without listening
to what I try to tell them about what I've learned.
So that's a quandary ...
If I try, once again, to sum up just what it is that
I think I have learned, for hope springs eternal and
all that, I think that this time, if I try to phrase
it freshly from the top, it would go a bit like this:
All we want is a logical system that helps us be smart.
Is that really so much to ask? When we turn to trying
to build such a system in the style of a computational
resource, which means looking for bits of work that we
can isolate and render computable, then there are just
some problematic obstacles that always seem to come up
and to block our way.
My experience with theorematics, with its prospective computational
facilitation, with trying to arrange for a computational assistant
that could catch the drift of inquiry and act like it understands
what it means to prove theorems like a real practitioner of that
art really does it -- well, I can tell you that the automated
implementation share of that has been largely an experience
with how logical systems crash. But there are some things
that one learns from carefully examining the different
ways that this can happen. By way of talking about
the salient phenomena that arise in these settings,
let me introduce a few bits of ad hoc terminology.
Let us describe a fixed set of axioms, definitions, predicates, and
rules or schemes for inference as a "fixed ontological genre" (FOG).
I my experiences charting, conducting, driving, navigating, piloting,
or scouting MOTA's and MOTO's through this or that FOG for lo! these
many years, I have learned to distinguish -- or learned that I ought
to distinguish -- the traits of the automated learning agent or the
running reasoning reagent that enable it to be what could be called
the "really intelligent logical system" (RILS) from the lack thereof,
that renders it, I suppose, a "non intelligent logical system" (NILS).
One of the things that I think I've learned in my sadder but wiser way
is this, that one of the qualities that really separates the RILS from
the NILS is this, the way that they behave when they finally meet with
that final contradiction to what they have been running full tilt like
they really and truly believe for many an eon of CPU secular existence.
And one of the characters that determines the fitful difference between
a viable conduct and a non-enviable conduct is this, that the RILS does
not allow itself to get trapped in quite so many FOG's as the NILS, but
acts as if it were somehow able to step outside of them and to move on.
Now, how is such a thing possible?
Exercise for the reader ...
Jon Awbrey
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤