Thursday, December 22, 2005
Lattice of ontologies
Function/structure descriptions
Paul Prueitt’s communication to ONTAC working group
Judith Rosen made a communication addressed to the ONTAC forum, and I forward this as she has not joined the forum.
In addition to some other remarks Judith said:
"However, where complex systems are concerned, such limitations equate to a requirement for radical oversimplification which causes the digitized version to diverge from the original system's behavior. The divergence typically increases as a function of time. Think about weather system computer modeling/forecasting; the forecast based on the computer model may be relatively accurate in the short term but is almost always inaccurate when the span of projected time exceeds 48 hours.
In order to usefully model complex systems with digital computers, it is necessary to constantly generate new models which follow the natural system over time. Robert Rosen compared this process to that of mapping a sphere (like planet Earth) using flat planes. It can be done, and very usefully, but it's important to know how that will change the nature of the original system. "
the rest of the discussion is at:
http://www.ontologystream.com/beads/nationalDebate/332.htm
This is a viewpoint shared by many in the complexity (natural complexity) community as well as other communities.
John Sowa just said to the ONTAC forum:
Paul
and Azamat,
This
discussion is going on far too long.
There are people in this group who have work to do. Pat Cassidy has
suggested that it be moved to the SUO list, and I agree.
Just
to respond to two points:
> Lev Goldfarb develops a wonderful new
approach on
> non-numeric models of ontological
realities...
Great! All those theories are in the lattice, which
contains
*every* theory without exception. [1]
> This is contrasted with the
"classes" of relationships
> between natural occurring concepts when
experienced
> and communicated within human
communities.
All
those theories are in the lattice. It
contains every
theory
without exception.
John
Patrick Cassidy said recently:
The general discussions of
basic issues on the ONTACWG list have
touched on topics that are
very much related to the question of how to
create effective knowledge
representations, but the immediate focus of
the ONTACWG is much
narrower, with a near-term goal to create
information artifacts that
will be useful to help people create
knowledge classification
systems and relate them to each other.
The
distinction between **related**
topics and **relevant** topics needs to
be recognized; many topics
are **related** to knowledge representation,
but **relevance** of
discussions depends on how well they serve to
advance a particular
goal. This distinction is part of the
charter of
ONTACWG (reference above):
"The focus of this
working group will be on the actual construction and
use of knowledge
classification systems. Discussion of basic principles
and theoretical issues will
be an important aspect of the collaborative
effort, but should in
general serve to address specific questions about
particular elements in one
of the knowledge classifications maintained
by the group or its
members. (2QPK)"
Accordingly, the discussion
on the general list should pertain to
specific Knowledge
Classification Systems(KCSs), development tools, or
specific applications of KCSs in existence or under development.
There are alternative KCS under development, such as my work, UDEF, and others. But Patrick and John are allowed to define (again in a narrow way that excludes real specific KCS) what a KCS is. A KCS becomes anything that is perfectly complaint to the W3C set of moving standards. This is deceptive and is game playing.
Respectfully, the specification of one precise meaning for a term in an ontology is an issue that is being ignored, by John, Chris, Patrick and the leaders of the ONTAC forum) because, the RDF use of URI (even with namespaces) has been taken in most cases to require that terms take on one meaning.
The alternative is to allow two things:
1) a reconciliation service to be instantiated that uses information about structural co-occurrence relationships (as discovered from a real time renewal of the "n-gram" analysis core to all (I think) semantic extraction (linguistic analysis or merely co-occurrence analysis) or (as edited, by humans, into evolving ontological models of processes and social discussion). (see RoadMap)
2) a late binding of an ontological construction (the actual specification of a "set of concepts with relationships, properties and attributes) AS A FRAME (but not as an OWL construction with "Logic"), and then the separate imposition (when appropriate) of some logical apparatus to form the type of finite state machine that W3C wants us to create. Here some lattice might be a Platonic form that helps lock in a situational standard for the purpose of a greater degree of data interoperability. This late binding can use a framework, such as Zachman's or Adi IF the framework helps real communities express what they want to express. (see papers on general Framework (gF) theory))
To give an example related to the use of lattices or with counting numbers. With counting numbers the "meaning" of the number 4 is set. However, having set this meaning abstractly, we find that human language exchanges allow an extension of the meaning of "4" in situations. Sometimes “4” means enough and sometimes “4” means not enough.
Without seeing a communicate-able model of the lattice of theories that John Sowa talks about it is hard to imagine any circumstance where a universal and sufficient ontological construction based on a single lattice (defined in some way) would find acceptance in the real world.
It is suggested that if John thinks it is a good idea then we should all accept the belief that he has something figured out that will be understood by the rest of use when we need to understand it. I do not think that such a position is consistent with science or with our cultural norms (at least up until just recently).
I communicating in the ONTAC forum because I wish to and because I feel that those in this forum are making decisions that needs to be more informed about alternative viewpoints.
[1] No John, the way that Lev develops his work specifically sets aside the notion of order as been artificial.