humanmarkup-comment message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [Elist Home]
Subject: [humanmarkup-comment] Knowledge Base Development Projects
- From: paul <beadmaster@ontologystream.com>
- To: categoricalAbstraction@yahoogroups.com
- Date: Sun, 07 Apr 2002 12:46:27 -0400
Kukahone,
Thank you for the
comment:
The best way to share a
paper might be to make the paper into a URL and post the URL. I can do
this and post into a scholars' section of the BCNGroup
The members of the
categorical abstraction forum will post comment about any scholarly paper that
is presented to the group. Our notions, derived from the BCNGroup Charter,
is on virtual collaboration and IP disclosure/protection. BCNGroup
membership is open as a means to support our work.
As you have indicated, the
concepts of emergence, and of semiotics, speaks about knowledge in a non trivial
manner.
But how is knowledge spoken
of today, in a way that is trivial? Is the knowledge representation
community and the community of computer scientists, generally missing something
important about the "ontological" difference between computer processing space
and the natural world? This difference is similar to the difference
between the concept of the number "three" and a pile of three stones.
The number three does not have quantum mechanical fluxuation as part of its
essence, whereas the pile of stones does (for example.) The number is
"non-complex", and the stones are complex.
We do not want this to
be a philosophical discussion, and so we define very clearly what the
business objectives are for challenging the way that "knowledge" is spoken
of today (by almost everyone - including Frank Sowa and most of the knowledge
representation community - Topic Maps, XML, RDF, KIF, Cyc Ontologies)
We are all,
everyone, looking for a programming/business-engineering unification.
Why? Because the social value of such a unification is larger than anyone
can imagine.
The
OSI/BCNGroup proposal is that the unification can be more easily made if
the natural science controls the computer science - and not the other way
around. We need to flip the current control situation in terms of how the
government and business regards the promise of computer science and the
application of computer science at the current paradigms in Information
Technology (IT).
In the near future, our
society simply must reduce funding of a AI mythology and the AI dream, and start
to fund, for the first time, a historically gournded paradigm where the
distinction between the ontology of the computer processing space and the nature
of natural systems is fundamental (as it is in the topic maps conceptual
model.) Why?
The IT mechanisms currently
available are wholly inadequate to address the complex time-critical problems we
are likely to face in the 21st century.
If we make this shift, then
things like a Cyber Defense Knowledge Base, Knowledge Base of COBOL
conversations, etc become feasible and will provide a high value to
society. Shining a light into the dark alley of the Internet is entirely
possible - but not with any of the current business propositions that come from
the Defense Industry.
By understanding the
natural science aspect OF computer science, then the problems of computer
science are by-passed in exactly the same way as the Zeno paradox, or the
Russell paradox is by-passed - (we simply quit talking about rational
constructs, like static pre-existing "scope" in Topic Maps that
are not productive - IN THE GENERAL CASE .)
But many of the thought
leaders actually are trying to reduce cognitive experience to formal constructs
- and in fact to formal constructs that are constrained by what Sowa calls
"first order logic", as if a certain interpretation of Frege, Peirce, Schroder,
Peano etc has in fact been able to reduce the phenomenon of mental experience to
a formalism. This is NOT the closed issue that Sowa would have us
believe.
Whereas the computer
science community may demand that such a reduction is possible, the natural
science community simply points out that
1) such a demand is
unreasonable based on specific principled arguments (related to stratified
complexity)
2) that there is simply no
evidence that current failures and limitations of knowledge representation will
be overcome - there is no example of an intelligence produced by a computer
process.
Real and natural emergence
is in fact what is missing. What we are left with, in my opinion, is the
potential for the development of semiotic control languages that treat the
computer processing space as a simple (but highly complicated) machine to be
described with machine based ontology. But as long as the
leading scholars in knowledge representation continue to push the AI Myth, then
capitialization of this potential work will fail to have a grounding in the
natural sciences, and the buisness community will continue to be mislead by the
promises of these standards.
We must reach into the
natural sciences to control a misunderstanding that is generating some
considerable error from the computer science community. A
reformed computer science will be more productive and will have greater social
value.
In a recent communication,
on a proposed Common Logic standard, to the topic maps community Dr. Sowa
said:
>One of the significant decisions was to choose a new
name,
>Common Logic (CL), for the proposed standard. The intent is
to
>reduce any bias toward the two starting notations, KIF and
CGs,
>and to emphasize the common basis in first-order logic, as
it
>was originally developed by Frege, Peirce, Schroder, Peano,
and
>many others during the late 19th and early 20th
centuries.
>
>In keeping with that decision, CL will be defined by an
abstract
>syntax, which specifies the major categories, such as
Quantifier,
>Negation, and Conjunction, without specifying any concrete
symbols
>for writing them. At the abstract level, even the ordering
is
>left undefined so that there is no bias toward a prefix
notation
>such as KIF, an infix notation such as predicate calculus, or
a
>graph notation such as CGs (or Peirce's original existential
graphs).
>
>Since it is impossible to write a purely abstract syntax, the
CL
>standard will also contain grammars for three concrete
syntaxes:
>KIF, CGIF (the CG interchange format), and traditional
predicate
>calculus (TPC) with a Unicode encoding of the commonly used
symbols.
>Each of those grammars will specify how the abstract categories
are
>mapped to the printable (or computer representable) symbols of
their
>notations. Any of the three concrete notations can be mapped
into
>any of the others by reversing the mapping from concrete to
abstract
>in one notation and then mapping from abstract to concrete in
the
>other notation.
>
>The standard will also contain a version of model theory
defined
>in terms of the abstract syntax. The model theory will
specify
>the truth conditions for any abstract statement, and any
conforming
>concrete statement in any syntax that is mapped from that
abstract
>statement would be required to have exactly the same truth
conditions.
>This requirement will ensure identical semantics for
statements
>represented in any concrete syntax that
conforms to the standard.
The facts may be that
scholars like Roger Penrose and Robert Rosen have laid out formal reasons why a
reduction of mental experience to a first-order logic will not be completely
successful. I will not repeat that arguments here except to give a
reference to my interpretation of this argument:
What I am proposing is that
the computer processing space (the Internet and all finite state systems
activity at any one time or over a period of time) is not a natural system in a
specific sense. The processing space is built to not have an influence
from any but one level of organization - and this organization is essentially
representable as a first order logic. This is the present computer science
that must be reformed.
So if one studies the
"addressable space" (in the language of topic maps) one can do this with a
standard like Common Logic, perhaps. The fact is that NO standard of
logic is being successfully used in such a way as to describe
the processing space AND be uniformly consistent with user
expectations. This failure can not be considered to be a failure of the
human species, but rather a failure of the reductionism in computer
science. Perhaps this was one of the points of the movie The
Matrix?
But perhaps this
description of the addressable space is only lacking a proper and universally
accepted standard. This is one issue.
The issue that has
been that we are all,
everyone, looking for a programming/business-engineering unification and
that many of the thought leaders are reducing cognitive experience to
formal constructs.
In doing this, we feel that a category error is made where by the
"formal system", which ever one you might be most happy with, is falsely
understand as if the formal system IS the natural system of business reality or
of the tacit knowledge of individuals.
***
We have specific approaches to gaining near term return on investment
based on the principle of separation of the artificial world of the computer
processing space and our understanding of the natural
world.
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [Elist Home]
Powered by eList eXpress LLC