Dear Leo, (01)
The good news is that you are right! (02)
> Humans already have an implicit commonsense ontology of the
> world. It's
> what we have acquired through our lifetimes, given perhaps
> some genetic
> predispositions or axioms. (03)
The bad news is that you are wrong! (04)
> 90-95% of that ontology is common to human
> beings. (05)
The reality is that we all have (sometimes subtly - sometimes not)
different ontologies of the world. (06)
The evidence is in every computer systems that has been built. If
we had the same ontology, you would expect that when we built
systems with that same ontology we would produce systems that were
(or at least very nearly were) compatible in their ontology. I have
yet to see any two systems that were compatible without a great deal
of work before the fact (and many of those failed). (07)
Regards (08)
Matthew (09)
> -----Original Message-----
> From: ontac-forum-bounces@xxxxxxxxxxxxxx
> [mailto:ontac-forum-bounces@xxxxxxxxxxxxxx]On Behalf Of Obrst, Leo J.
> Sent: 18 May 2006 23:23
> To: ONTAC-WG General Discussion
> Subject: RE: [ontac-forum] Problems of ontology
>
>
> Folks, John and Azamat,
>
> Humans already have an implicit commonsense ontology of the
> world. It's
> what we have acquired through our lifetimes, given perhaps
> some genetic
> predispositions or axioms. 90-95% of that ontology is common to human
> beings (this is anecdotal and cannot be scientific at this time). What
> the ontology engineering effort is about (among other things) is to
> formalize that knowledge to enable machines to interpret that model
> (lattice of models, if you will) and thus enable software to take
> advantage of human semantics. To draw inferences comparable to human
> inferences (without machine understanding; this is a syntactic
> operation; there is no AI doing this), given that same knowledge,
> approximate as it is.
>
> Commonsense and scientific theories figure in that mix.
> Philosophophical formal ontology theories figure in that mix. We need
> to put the best human theories into that mix. Theories of
> mereology or
> mereotopology aren't scientific theories in the usual sense, but they
> are theories. Formal semantics of natural language largely falls back
> on refined intuitions about interpretations of natural language. That
> is good, because what else is there: there is only
> generalizations from
> data and the reflections of those that individual human beings have
> regarding their analyzed notions about their own language usage.
>
> Yes, we can have human descriptions of those ontologies.
> Largely that's
> what philosophy has done for the past 2000+ years. Or we can put those
> descriptions into a logical form (yes, using a formal tool that humans
> may have invented or that has been there all along -- this is not an
> appeal to idealists, nominalists, realists, quantum information
> scientists and their individual perspectives and theories, but intends
> to transcend those).
>
> And then that logical form can be syntactically utilized by machine
> inference engines that know about the logic the content is formalized
> in, and can apply stepwise the syntactic proof or semantic
> tableaux (or
> other methods) methods to perform machine reasoning (deduction,
> induction, abduction, your-favorite-uction) on those symbols. Thus
> enabling machines to work much more closely to the human conceptual
> level than they can now.
>
> Throwing up your hands in either direction -- 1) we can have no
> ultimate common human understanding, and 2) we cannot formalize any
> common human understanding that is usable by machine -- is to me,
> something to be ascertained by way of empirical experiment and NOT
> ascertained ex nihilo, by philosophical or idealogical argument.
>
> I believe there are limitations to the scientific method, but I don't
> think these figure here. This is standard science.
>
> Leo
>
>
> _____________________________________________
> Dr. Leo Obrst The MITRE Corporation, Information Semantics
> lobrst@xxxxxxxxx Center for Innovative Computing & Informatics
> Voice: 703-983-6770 7515 Colshire Drive, M/S H305
> Fax: 703-983-1379 McLean, VA 22102-7508, USA
>
>
> -----Original Message-----
> From: ontac-forum-bounces@xxxxxxxxxxxxxx
> [mailto:ontac-forum-bounces@xxxxxxxxxxxxxx] On Behalf Of Azamat
> Sent: Thursday, May 18, 2006 5:52 PM
> To: ONTAC-WG General Discussion
> Subject: Re: [ontac-forum] Problems of ontology
>
> >> AA> For without a common standard ontology as a common code of
> meanings
> >> > and rules, there is no base and foundation for the whole
> enterprise
> >> > of ontological semantic technology, or intelligent
> applications...
> >>
>
> >>JS > No! Absolutely not! That is the ultimate goal, not the
> starting
> >> point. The best we can say is that every attempt to design a
> >> unified theory of everything has failed. Yet somehow the genus
> >> Homo has survived for about two million years without it. Perhaps
> >> we may have a better foundation in another million years or so.
>
> John,
>
> No need wait for so long time, since such an ontological grounwork
> begun at
> the moment humans started to systematically meditate about the
> environmental categories such as being, reality, existence, cause,
> time,
> space, life, etc. As a result, any substantial doctrines, religious,
> ideological or political or social, are pervaded with ontological
> ideas.
> Any science, formal or material, presupposes ontological assumptions
> about
> its universe
> of discourse. For every general scientific theory is inherently
> ontological,
> and other way round, every significant ontological theory is
> scientific.
> Inside, ontology is a general science as much as any science is a
> particular
> ontology. Implicitly and explicitly, modern science and
> technology have
>
> ontological inputs and outputs.
>
> Currently there are several knowledge domains closely affiliated with
> ontology: science, mathematics, logic, and semantics. Using them as
> conceptual tools and instruments, ontology provides the most general
> theories about reality; with its referent and scope, reality; goal,
> underlying uniformities and patterns and laws in the
> universe; method,
> postulation; truth type,
> ontological; and role, integrated world views and computable
> models for
> intelligent technology.
>
> So neither theory, scientific or mathematical, nor formal logical
> languages
> nor natural languages can do without an ontological background,
> enhanced
> with mathematical axioms, scientific formulas, and semantic
> assumptions.
> Accordingly, without ontological groundwork, only formal logical or
> pure
> semantic or mathematical or scientific or linguistic analysis would be
> incomplete or even misleading. This also goes for your idea of the
> lattice
> of theories, which to be viable and appealing should be based on a
> single
> unified ontology theory, which rudimentary axioms and truths and laws
> and
> principles and rules will make the fundamental assumptions
> for the less
> general lower levels: domain-independent theories, domain models, and
> specific
> descriptions.
> Lack of such understanding is a major obstacle to the success of large
> ontological projects as the semantic web, CYC, etc.
>
> With due respect,
> Azamat Abdoullaev
>
>
>
> >
> > ----- Original Message -----
> > From: "John F. Sowa" <sowa@xxxxxxxxxxx>
> > To: "ONTAC-WG General Discussion" <ontac-forum@xxxxxxxxxxxxxx>
> > Sent: Thursday, May 18, 2006 1:57 AM
> > Subject: Re: [ontac-forum] Problems of ontology
> >
> >
> >> Azamat, David, and Eric,
> >>
> >> First of all, the view I presented is definitely not a rejection
> >> of "the ultimate goal of science". It is, in fact, a recognition
> >> of the difference between the process and the goal.
> >>
> >> JS>> The view of science as a unified body of theory is totally
> >> >> unrealistic. Even physics, the hardest of the hard sciences,
> >> >> is a hodge-podge of thousands of mutually inconsistent models
> >> >> for each area of application. As Bundy said, there is no such
> >> >> thing as a unified model of everything.
> >>
> >> AA> What i found so disturbing about your position, it is not
> >> > its partiality, but rather its apparent rejection of the ultimate
> >> > goal of science led by ontology, to set an all-comprehensive and
> >> > unified description of the world.
> >>
> >> My basic claim is that science is not yet finished. Science is
> >> first and foremost a methodology for asking and answering questions
> >> about the universe. And every research scientist will tell you
> >> that whenever they find the answer to any important question,
> >> they discover that it always leads to even more questions.
> >>
> >> And what on earth do you mean "led by ontology"? Aristotle
> >> explicitly put metaphysics *after* physics because it's
> >> impossible to do any decent work in ontology without first
> >> having a great deal of empirical evidence. Without empirical
> >> data, the only subjects you can work on are pure uninterpreted
> >> logic and mathematics. They're useful tools, but certainly
> >> not capable of "leading" science.
> >>
> >> AA> And the high goal of science...
> >>
> >> We don't disagree about the goals. My only point is that we
> >> still have a long way to go before we reach them. Perhaps
> >> we might get there in a few more centuries, or perhaps it may
> >> take billions of years. Nobody knows.
> >>
> >> AA> ... the task of ontology is to give us the overall structures,
> >> > uniformities, patterns, laws and constraints within which all
> >> > the many changes in the world take place...
> >>
> >> I don't disagree. My only point is that it might take us millions
> >> of years to get a complete answer. We certainly don't
> have anything
> >> remotely resembling a final solution today -- or even a
> good outline
> >> of a final solution.
> >>
> >>
> >> DE> That's the best description of the enterprise collection of
> >> > legacy systems -- what we're tying to inter-operate on, right?
> >> > -- that I've seen in a long time.
> >> >
> >> > While there may in fact (in the class-room) be something called
> >> > "computer science," such purity is very hard to find in legacy
> >> > systems land.
> >>
> >> Unfortunately, *every* system becomes a "legacy system" the instant
> >> it is released to the customer. That is when the
> engineer's dilemma
> >> kicks in:
> >>
> >> Customers never know what they want until they see what
> they get.
> >>
> >> That causes version 1.0 to be replaced by version 1.01, 1.02....
> >> And every version is incompatible with every previous version,
> >> with every subsequent version, and with every piece of software
> >> produced by any other vendor.
> >>
> >> DE> I wonder if there is any correlation between how valuable
> >> > a system is and how chaotic it is.
> >>
> >> Every system becomes more chaotic with each release. Windows NT
> >> version 3.5 was probably the cleanest. That's when Bill Gates
> >> declared that the support for the display had to be brought into
> >> the kernel for NT 4.0 -- then the bugs really began to multiply.
> >>
> >> Today, Windows XP is bad enough, but Digital Rights Mismanagement
> >> (DRM) puts chaos directly into the kernel of the operating system.
> >> If you want to preserve even a modest amount of sanity, I suggest
> >> that you buy your last version of Windows in 2006 before Vista is
> >> inflicted upon us. For 2007, do not think of using any software
> >> that incorporates DRM. (For evidence of this point, look at the
> >> iPod by Apple. Sony failed because they tried to implement DRM
> >> in their device, but Apple wisely ignored DRM.)
> >>
> >> EP> All my thoughts and effort toward mapping and merging upper
> >> > ontologies have been centered on the creation of a merged lattice
> >> > of types and and a merged lattice of relations. For my purposes,
> >> > this is where I seem to be getting value. Theory boundaries seem
> >> > to me to be more arbitrary and less useful stuff with which to
> >> > integrate databases. I'm still not sure if we actually differ
> >> > on this point. I'm suspecting that we agree.
> >>
> >> I agree that working with a lattice of types and relations is
> >> important. But as soon as you start writing axioms for those
> >> types, contradictions begin to proliferate. I like to quote
> >> Peirce's remark (CP 4.237):
> >>
> >> It is easy to speak with precision upon a general theme.
> >> Only, one must commonly surrender all ambition to be certain.
> >>
> >> It is equally easy to be certain. One has only to be
> >> sufficiently vague.
> >>
> >> It is not so difficult to be pretty precise and fairly certain
> >> at once about a very narrow subject.
> >>
> >> The first point summarizes the futility of a trying to develop
> >> a precisely axiomatized ontology of everything, but the second
> >> point holds out hope of having a decent, but vague overview.
> >> That's why WordNet is so useful -- it's vague, and it has very
> >> few axioms. Unfortunately, it still has contradictions.
> >>
> >> The final sentence corroborates the point that Lenat makes about
> >> Cyc -- the most important inferences are in the microtheories,
> >> which are, in Peirce's terms, "very narrow subjects" about
> >> which it is possible "to be pretty precise and fairly certain".
> >>
> >> Summary: I agree with Azamat about the goals of science, but I
> >> don't believe there is any evidence that we are even close to
> >> reaching a final theory of everything. For supporting evidence,
> >> I highly recommend the following book (which assumes a good
> >> mathematical background, by the way):
> >>
> >> _The Road to Reality: A Complete Guide to the Laws of the
> >> Universe_ by Roger Penrose, Alfred A. Knopf Publishers,
> New York.
> >>
> >> For an estimate of how close we are to a final theory of physics,
> >> read the last Chapter 34, "Where Lies the Road to Reality."
> >> Penrose believes that there is likely to be a major upheaval
> >> sometime in the 21st century that will be at least as far reaching
> >> and perhaps even more disruptive than the revolutions caused
> >> by relativity and quantum mechanics in the 20th century.
> >>
> >> That shows how far we still have to go just to find the basic
> >> equations for the foundations of physics. And even if we knew
> >> the equations, they still wouldn't tell us what approximations
> >> are necessary in order to make them applicable and solvable for
> >> any practical problem.
> >>
> >> So much for the hardest of the hard sciences. When it comes
> >> to economics, sociology, psychology, and their applications to
> >> business and politics, we have *nothing* that can even come
> >> close to 19th century physics. We have a lot of practical
> >> experience about "human nature", but *nothing* so far that
> >> remotely resembles even the bare beginnings of a vague theory.
> >>
> >> For the foreseeable future, the best we have are lots of usable
> >> piecemeal results on narrow subjects, but no general theory of
> >> everything. That doesn't mean we should give up work on
> >> fundamental science, but there is no end in sight.
> >>
> >> Meanwhile, let's make use of what we have.
> >>
> >> John Sowa
> >>
> >>
> >> _________________________________________________________________
> >> Message Archives: http://colab.cim3.net/forum/ontac-forum/
> >> To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
> >> Subscribe/Unsubscribe/Config:
> >> http://colab.cim3.net/mailman/listinfo/ontac-forum/
> >> Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
> >> Community Wiki:
> >>
> http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCo
ordinatin
gWG
> (010)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config:
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatin
gWG (011)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config:
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (012)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config:
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (013)
|