Early on in this discussion John used the example of terminologies “cutting up a continuous range of variation into a discrete set of categories” and offered examples of:river, stream,
creek, brook, rivulet. He then noted that:
JS> Such terms can make up a very useful terminology.But they aren't part of an ontology
until you develop some theory about what they might denote.
Later John clarified the
use of theory in the sense we have from Logic as a deductive closure over
axioms.I know this topic has come up
before, but I sometimes feel that we are jumping back and forth between theory
in the sense of science theory (QED was cited as an example) and theory from
the point of view of Logic.
I see both as relevant and
like some of the distinctions that Nicola Guarino made back in the intro to Formal
Ontology in Information Systems (1998). The way I interpret his formulation
is that a formal ontology product is a conceptualization
for a Domain (C for D) by means of axiomatic commitments (K) expressible in a
Language (L) using some Model. Viewed this way ontologies start from “Analysis
and Conceptualization”.
John’s example was the
cutting up of some water domain into various categories (rivers, steams etc.) that
often come to have a term associated with it. We might have an application dealing
with floods in which these distinctions are
important. To start on a quality ontology for such an application it should be
able to make meaningful statements about what exists in its focused domain.So river and stream ideas might be organized
along a hydrological theory of what brings water collections into existence and
what natural processes, such as downhill streaming flows, they follow. The
hydrological concepts are more basic and underlie the real world phenomena at
the river-stream level. For many applications they will therefore organized
things.
Thus :
water within a river (or other watercourses) is
generally collected from precipitation through surface runoff, groundwater
recharge, springs, as well as the seasonal release of stored water.
Stored water may include what is held by dams
but usually is from natural ice and snow packs (e.g., from glaciers).
This conceptual theory
specifies various classes of real objects (e.g. snow pack) and process (e.g. spring
runoff), and relations (e.g. precipitates, thaws) that we assert applies among
instances of such classes, as well as relationships among such classes and their
instances.
Now all of this
conceptualization may precede the formalization into axioms, but for IT
applications we need to formalize these axioms in a language that on the one
hand faithfully reflects this conceptualization and on the other can be processed
by applications.
"And therein lies the problem: why do we have to choose the best one devoid
of context? "Horses for courses" as they say back in the UK. There is no
universal best one :-)"
Agree.
Each good definition, as well
as classification, has its strength and weaknesses.
It comes from different ways we define words
or classify things. We construct definitions by reference to:
1. genus and differentia
2. causes (genesis, origin, material cause,
formal cause, final cause, or productive cause)
3. principal features
4. functions
5. dichotomy or division
6. component parts
7. extrinsic signs
8. purposes or interests, etc.
On the positive side, there are two broad types
of ontological definitions/classifications: real definitions and
nominal definitions. The nominal one is mostly arbitrary, relating
substance with accident, while the real definitions/classifications tend to
reflect/represent the nature of things as adequately as possible. I
believe, it could be used as a criterion of best choice.
"Taken as pure and abstract knowledge, Ontology is formulated as
different as:
the science (account)
of entity (or being) in general;
the
knowledge of the most general structures of reality;
the theory of the kinds and
structures of things in every domain of reality;
the study of entity types and
relations;
the most general theory
concerning reality, being, or existence;
a collection of absolute assumptions;
the study of change;
the science of all possible worlds
and everything conceivable;
the study
of semantic values of natural and formal languages and ontological
commitments about the world
In the context of computer
science, information and communication technologies, an ontology is reckoned
to be:
a set of generic or
philosophical concepts, axioms, and relationships for domain ontologies
(IEEE SUO, 2000, 2003);
a taxonomy of
world terms/categories comprising definitions, hierarchical relations, and
formal axioms (Mizoguchi, 1998);
a
set of definitions of classes and their relations, as well as individuals
and their properties (OWL 2004; 2006);
a catalog of the types of things
(representing the predicates, word senses, concept and relation types of
some formal language) organized by the class-subclass taxonomical relation
(Sowa, 1997; 2000);
metadata schemas
with machine processable semantics (Horrocks, 2003);
content theories about the kinds
of objects, their properties and relationships possible in a certain
knowledge field (Chandrasekaran, Josephson, and Benjamins, 1999);
the total of a taxonomy and a
set of inference rules or a document (or file) formally defining the
relations among terms (Berners-Lee, Hendler and Lassila, 2001)"
Try and choose the best one.
And therein
lies the problem: why do we have to choose the best one devoid of context?
I. "Ontology is a general theory about the
world, its domain, entities and relationships."
II. "An ontology is a general theory about some
aspect of the world, its subdomains, entities and relationships."
III. "A formal ontology is a formal theory of
some aspect of world, its subdomains, entities and relationships."
My recommendation for anyone who is trying to define
anything is to check a good dictionary for an independent opinion.
Following is the definition from the closest dictionary at hand,
_Merriam-Webster Ninth Collegiate_:
1. a branch of
metaphysics concerned with the nature and relations
of being.
2. a particular theory about the nature of being
or the kinds of existents.
I don't know who
wrote those two definitions, but they're as good as any and much
better than most. The editors of the best dictionaries usually have
associate editors for various fields. The person who wrote (or
reviewed rand revised) that definition was probably a philosopher who
was knowledgeable about the field.
In English, the word 'ontology'
without a preceding article refers to the branch of philosophy.
With an article or other determiner, such as "an ontology",
"Aristotle's ontology", or "Kant's ontology", it refers to a specific
theory.
My suggestion is to adopt the distinction from M-W.
In the discussions in this forum, we're usually talking about specific
theories. That means all of them are variations of M-W
definition #2.
I also recommend an adjective, such as 'general' if
it has a broad scope. If it has a more narrow scope, I would add
a qualifier, such as 'medical', or a name, such as 'XYZ Corporation'.
Another adjective would be 'formal' if the definitions are stated
in some version of logic or mathematics. By combining the
adjectives, you could talk about a formal general ontology or an
informal medical ontology.
If you leave out the adjective
'formal' or 'informal', it avoids making a commitment about whether
the terms are stated in some version of logic. The default
assumption is that they're not, but it leaves open the option of a
future revision and extension that defines some or all of the terms in
some version of logic.