[ontolog] So what is an ontology?

["Brian (Bo) Newman" <bonewman@xxxxxxxxxxxx>]

Esteemed Colleagues, 

Thomas Lee asked if anyone could suggest some readings that go beyond the
introductory level. While, due to its obscurity, it might not be of much
help to Thomas, my favorite is a little known 1948 work by Paul Glenn
titled "Ontology, A Class Manual in Fundamental
Metaphysics."  One of the reasons I like this book, and the
reason I mention it, is because, as it was written before the age of
computers, it helps me focus on the foundational nature of this field of
study.  Whereas epistemology is the "science of knowledge"
-- and by-in-large a study of the abstract -- ontology on the other hand
is the "science of being," and far more connected to the real
world in which real people, organizations, and even computers (all forms
of agents) communicate, collaborate, and attempt to work together to
solve real problems.  

While this is clearly stating the obvious to this audience, I believe
that agent's ontologies -- their 'world-views' -- lie at the heart of all
interactions.  Only if the operative concepts and/or ideas have
'being' and some degree of consistency for the agents involved, can there
be meaningful communication, iteration, or collaboration. The same is
true for this group and what we might hope to accomplish together. 

Individually we each have our own ontologies, and within our individual
views of the world, as evident by the fact that we have chose to be a
part of this group, we all seem to have this concept of something called
"an ontology." However, I don't think it would come as a
surprise to any of us that we probably don't all have the same
understanding of that concept. I would think that developing a shared
understanding of just what an ontology is, and is not, would be one of
the first things we would want to address. 

Having said that, I will share a few of the ways that I think of
ontologies. 

The relationship between epistemology, ontology, taxonomy, semantics, and
meaning can be stated be saying that: Epistemology tells us what how
ideas and concepts can exist. Ontology tells us what ideas and concepts
do exist.  Taxonomy provides as with a means for developing
classification schemes for elements of an ontology in its
representational form. Semantics provide a means for resolving the
ambiguities of meaning resulting from less than perfect representation.

Agents have ontologies and ontologies only exist within the
"mind-space" of agents, be they individual agents (people),
automated agents (technology), or collective agents (groups,
organizations, businesses, etc). 

Ontologies are more than a collective set of definitions. Ontologies are
the synergistic nexus of previously retained knowledge (concepts and
ideas), and the relationships between those concepts and ideas. 

Ontologies deal with reality, but not with the material objects of
reality (which is the subject of science). Instead the deal with the
properties of being.  Ontologies define what that can exist
(have being) within an agents world, the nature of those potential
realities (abstracted characteristics), and the relationships between
those things.  

Ontologies are held by agents as retained knowledge, and as such can be
explicit, implicit, or tacit depending on the knowledge retention
capabilities of the agent.  In those cases where the agent is
capable of creating or assimilating "new" knowledge or
"forgetting" previously retained knowledge, ontologies are
subject to change.  Ontologies represent a part, but not all of an
agents retained knowledge.

Ontologies are seldom if ever applied in their entirety, but rather as
appropriate to a given situation (operative ontologies). 

Of course, this is just a reflection of a portion of my personal ontology
...  

Who would like to go next?

- Bo Newman - 

--
To post messages mailto:ontolog@xxxxxxxxxxxxxxx
An archive of the [ontolog] forum can be found
at http://ontolog.cim3.org/forums/ontolog