[33]                               home                          [35]


4/29/2004 8:36 AM


Key questions on Common Upper Ontology


Key questions on Common Upper Ontology -> .


In many of these beads we have addressed the issue of ontology use.  We have had to struggle with the use of language adopted by Semantic Web advocates and by computer science.  Words like “semantics” are characteristically over loaded and are often treated as if one set of meaning does apply, when the actual standards work is far far more limited than what many natural scientists feel comfortable with.


John Sowa made a comment on the actual nature of standards work at the W3C as being about how computers exchange bits, not about the science of natural systems. 


Dr. Sowa wrote:


The discussion about ontology creates a lot of confusion, partly because "ontology" is very seldom defined, and when it is, the definitions are so vague (e.g. "specification of a conceptualization") that they mean practically nothing.


Since the applications we are currently considering are related to computing, I suggest that we replace every occurrence of the word "ontology" in these discussions with the term "program specification" or just "program spec”


Sowa’s work is introduced into the BCNGroup’s Glass Bead Game in this bead.  


As is discussed in [33] the separation of


a)        programming specs

b)       from the notion of the meaning (or causes) of something in the context of human communities


is essential to ending the confusions caused by artificial intelligence and some semantic web language.  Our position is the CoreSystem’s project is far more optimal for program specification than the W3C standards committee process could ever come up with.  The W3C is just too invested in existing confusions and agendas (we respectfully claim).   We urge the intelligence community to adopt the CoreSystem and the Knowledge Sharing foundation concept and to immediately deploy a system that examines the nature of social discourse as a means to address the memetic battlefield.  We note that In-Q-Tel invests in companies like Intelliseek Inc.  Our position is that these companies poorly use Anticipatory Web thinking.  There is a danger in the incomplete use of knowledge technology. 


We note that our several proposals to DARPA, NIST, In-Q-Tel (CIA), and NIMA were dutifully developed and submitted.  In our NIMA submission we were deemed fundable by peer review but the funds for NIMA’s Glass Box ontology development were given to Cyc Corp and we were told that no funds were available to develop the Knowledge Operating System, based on categoricalAbstraction, eventChemistry and the Actionable Intelligence Process Model. 


Machine ontology, (also called formal ontology) in the sense that Sowa is suggesting, is about how the computer works.  Natural ontology is quite different in a precise way that I talk about in Chapter Two of Foundations of Knowledge Science, “Is Computation Something New? ”.  In this on-line book, in this specific chapter, I express the position of the BCNGroup that science must look across the table at the business people and the computer scientists and say”


Some scholars make a distinction between a formal system and a natural system and hypothesize that no natural system is in fact computational in the formal sense (Rosen, 1995). These scholars suggest that the forces shaping the evolution of a natural system are not computations.


In saying this we are not trying to sale anything to anyone.  We make a principled observation about natural science and about the nature of complex natural systems.


So the concept of “ontology” has precisely two contexts, one in the context of mathematics and computer science and one in the context of a natural system such as a vineyard or a microfarm.  I use the phrase “formal ontology” to refer to mathematics and computer science and the phrase “natural ontology” to refer to natural systems.  Formal ontology is abstraction and natural ontology is not. 


The formal ontology existing as part of a computer program is an abstraction, even if the state transitions are real natural events.


The confusion develops when formal ontologies are used in machine translations or in the publishing industry.  The confusion is not helpful, and yet is entrenched in our social discourse and in our funding decisions.


The Dublin Core is a primary example. 


The Dublin Core Metadata Initiative is an open forum engaged in the development of interoperable online metadata standards that support a broad range of purposes and business models. DCMI's activities include consensus-driven working groups, global conferences and workshops, standards liaison, and educational efforts to promote widespread acceptance of metadata standards and practices


This system of “subject matter metadata” is an extension of the notion that an exact description of “the nature of reality” (ontology in its historical meaning), can be developed and used as if one where talking about a formal description.  Protégé and Cyc Corp make similar claims on the term “ontology”.  Their claim is that the distinction made by myself and others (as discussed in Chapter Two) can be ignored. 


The BCNGroup’s observation is that there is no productive reason to ignore the natural science and maintain the confusion caused by decades of heavy federal investments in artificial intelligence.  The conflict is between those whose academic and industrial jobs depend on the confusion, and those who need to use computers in a more profound fashion than what has been up to now possible.  It is a natural conflict similar to the conflict between the electrical direct current industry (supported by Edison) and the electrical alternating current industry (supported by Westinghouse and Telsa) in the early part of the twentieth century. 


The transition between DC technology and AC technology involved the blocking of AC by the economic and intellectual power of the DC industry.  But the truth was that most of the DC engineers found a comfortable home in the AC industry.  They just had to leave behind the antagonism towards AC caused in large part by the ego of Edison and the perceived self-interest of Edison’s business partners.  The parallel; between


(a)    the knowledge sciences and AC and

(b)   the Information Technology and DC is developed in some beads.


As observed in the previous bead, the work on knowledge bases for publishing is important work and has lead to beneficial outcomes. Medical ontology developed in Protégé is very useful in medical education, for example.  However, in the intelligence community program managers are influenced to make investments in AI R&D such as the Cyc Corp “ontology” when the claims of Cyc Corp have been found wanting. 


{ 4 } , [ 5 ]  ( 33-4 )


John Sowa made a presentation, QQAM Summer Institute on Cognitive Science (June 30, 2003) on alternatives to Cyc and said


Cyc project started in 1984 by Doug Lenat.

Name comes from the stressed syllable of encyclopedia.

Goal:  implement the commonsense knowledge of an average human being.

After $65 million and 650 person-years of work,

600,000 categories
defined by 2,000,000 axioms
organized in 6,000 microtheories.

But it cannot compete with a 10-year-old child.


So Cyc Corp, and other heavily funded Corporations, continues to talk about the potential of Cyc ontologies.


They continue to ignore the principled arguments that logics developed on the atoms of tree like graph structures and first order predicate logics can not work in the general case.  And then they, Cyc Corp and others, continue to get large federal awards based on the assumption that eventually the hard AI position will be shown to work in normal circumstances. 


The conter-arguments made in Chapter Two follow those of Sir Roger Penrose and others.  The others includes groups who are not addressed in Sir Roger’s book, groups like The Einstein Institute and academics within the fields of ecological psychology. 


It is a big deal!  The collision between formal ontology and machine ontology is not optimal in terms of value to society. 


The standards process has largely hijacked the discussion and the resources available and preserves the AI mythology largely in order to maintain the importance of several hundred people.  A number of scholars in the natural sciences are able and willing to continue, but lack resource to even meet and have a conference.   The individuals supporting the Semantic Web concept most often have government or industry employment that is long term and which supports their advocacy.


But this is pure ego.  The self centered-ness of Industrial players works with individual ego to produce the current blindness to the many issues that have been raised, such as by the authors of the original Topic Map standard.   The standards process is now designed to waste the effort of those who would bring up the deeper issues related to natural ontology.


But as in the Edison-Westinghouse collision, the shift to AC-IT will occur and many of those who now oppose even the discussion about natural ontology will find a new and rewarding home. 


We are early in the process, and feel the economic pain.  Individual memberships ($75) and donations to the BCNGroup help.