Re: [ontolog-forum] Ontologies and individuals

[Amanda Vizedom <amanda.vizedom@xxxxxxxxx>]

Sandro,  

I'm struck by some aspects of your question. My first reaction is a rather simple "of course" and "why not?" That is, I am curious as to why one would think that an ontology could not include individuals.  I can't think of anything about ontologies in general that would suggest this.  Of course, not everyone accepts the idea that "an ontology is a theory about a conceptualization." Personally, I think some ontologies are, or contain theories about conceptualizations -- one *can* model a conceptualization -- but this need not be useful.  Rather, the ontological modeling toolkit, like any other modeling medium, can be used to model a wide range of things.  

Second, I was struct by this phrasing "... usually refer to concepts; not to individuals."  The assumption of disjointness here is what seems odd to me.  Certainly, if you define "concept" to exclude individuals, it will be hard to see how individuals fit in a concept-based model.  I'm also aware that certain sub-communities of ontologists use the English word "concept" for what other ontologists call "class" or "type" or "kind" or even "collection".  I have seen this in particular biological ontology communities. Here, I think what is going on partially from the heritage of biological taxonomy, in which the taxonomic/ontological structure is a hierarchy of classes, and is separate from treatments of classification of individuals into those classes. Furthermore, the traditional subject matter is defined with sufficient narrowness that reference individuals (e.g., earth) are implicit.

Well, I'm only speculating on the cause. But I've observed the pattern, especially as it tends to generate clashes when ontologists who have worked only within such conventions join teams working in other domains with no such convention, or when there is a need to reuse or align bio ontologies that follow this convention with others (bio or other domain) that do not. 

Beyond that, I agree with Alex's and John's points regarding special case individuals. In fact I am racking my brains for examples  of real, in-application ontologies I've worked with that did not require such individuals.

It may be possible to create and use a very partial, very shallow ontology for some limited domain that does not contain any individuals, but I cannot think of an example in which this could be the case at the same time as the ontology has any formal depth. That is, if one takes a hard line that *no* individuals be included, a consequence seems to be that for many of the important classes and properties in the ontology, there will be little explicit semantics; defining and distinguishing features of those classes and properties will not be specified.  This reduces the usefulness of such an ontology for many application, as well as maintainability. The more major concept definition is left implicit, the more interpretation will vary by user, over time, etc.

In any case, it is the norm IME to have at least some individuals. There is lots of variation regarding degree of inclusion of individuals.  Reference individuals -- i.e., those in relation to which other concepts are defined -- are commonly included. Beyond that, it varies by the nature of the application and  on choices between different abstract models of things like numbers (e.g., is a real number an individual? Is it a set?  Are groups defined? Or only the sets of members of groups?)

In the end, there is not a crisp line but rather a continuum between ontologies and ontological knowledge bases.  I see reasons for making practical choices of what to include in particular cases, but I don't see any principled reason for excluding individuals.

Amanda

On Mon, Dec 10, 2012 at 3:25 PM, John F Sowa <sowa@xxxxxxxxxxx> wrote:

Sandro and Alex,

To be general, an ontology would emphasize types rather than
specific individuals.  But there are many categories that
contain unique individuals that must be considered.

SRF

> There is a question we have been discussing in our group for
> a long time, and we would be glad to have your opinions on it:
> Can ontologies really contain individuals?

AS
> But even [in general theories], there are a lot of individuals

> for ex. in arithmetic (0, 1...).

Yes  But there are other reasons for distinguishing special cases.
Astronomy, for example, makes a special case for one star:  our sun.

Geographical information systems make a special case for one particular
planet:  earth, its continents and coordinates.  The earth's moon is
also treated as a special individual.

Legal systems are always specific to one particular country, and they
also have to distinguish changes over time.  The subdivisions of each
country may also have legal systems that have quite a bit of divergence.

As another example, the kilogram is defined in terms of one specific
individual:  a cylinder of platinum-iridium stored in a vault in France.
There are recommendations to redefine the kg in terms of Planck's
constant, but no official decision has been made.

John

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)