October 3, 2005
Center of Excellence Proposal
à
Communication to a Semantic
Web standards related e-forum
A focused position statement
on the notions that has been briefly discussed here regarding the artificiality
of semantic web ontologies is at [bead 188 and 189]
Communicated to http://protege.stanford.edu/community/subscribe.html
In preparation for a organizational meeting on semantic web interoperability, Dr Patrick Cassidy, at Mitre, developed a 62 page PowerPoint presentation. On page 54 he wrote:
For the Skeptical
Help the process:
It will be useful to have a set of use cases or scenarios that would
provide a practical challenge for the developers of integrated technologies
such as the Common Semantic Model. What
would satisfy the variable ?this :
“If you can do ?this, I will be convinced that a Common Semantic Model is valuable.”
Having talked to Patrick a few times, and I think I know (rhetorically) what he does not want as the use case.
Suppose that the use case was one in which an emergent feature plays a critical role in the function of an event.
Suppose that ?this is a phenomenon that needs to be represented as a set of syntagmatic units { < a, r, b > }, with some constraints placed on these units so that they are within the W3C RDF (Resource Description Framework) and using descriptive logics or predicate logics.
One has to be careful here. I feel, as do others in my community, that an ontological construction (which I defined to be a functional model of an aspect of reality) can be developed that fulfills the use case involving significant emergent properties. But Patrick wants to approach this as something subsumed what his community has called “upper ontology”.
Upper ontology is supposed to be common to all ontological models, and by common it is meant “a subset of”. He also wants to define and support the use of the RDF/OWL W3C standardization and the use of first order predicate logics. I say that “he” wants to and look to him to make corrections to how I have represented his viewpoints.
There are uses for the RDF/OWL W3C supported standards. The question is about extension from special cases where something does works, to the more general case.
The extension question is a serious one, and has to do with whether these standards will go all the way in delivering “semantic interoperability”. My group claims that the issues related to modeling emergence are critical. Patrick seemed to me to want to say that emergence is not definable or is not relevant or something else that allows him to avoid the issue that functional general-purpose ontologies are not being found working in any real or complete sense (anywhere). What we have are a lot of computer scientists pulling the wool over other people’s faces.
The notion of emergence is critical if one is to go only a little way in doing what is called a “merge of two ontologies”. Of course one can hand construct instances were a, technically defined notion of, ontology merge can be shown to occur.
But does this merge have a full correspondence to a complex social reality (the one we live in).
I say that many natural scientists would say “no” the W3C standards couldn’t produce operational ontology that supports functional ontology merges. The formalism is just not strong enough, or too strong – depending on how one thinks about relationships between things like set theory and fuzzy set theory, or category theory and rough sets.
Paul Prueitt
Taos, New Mexico