On 12/12/2012 11:09 AM, John F Sowa wrote:

Sandro, Bill, Richard, Hans, Alex, Joel, and Hassan,
There is a big difference between guidelines and requirements. In most
cases, a guideline that permits exceptions is more useful than a strict
requirement or prohibition. Ontology is still a research area with many
unsolved, poorly defined, or even unknown problems. We cannot predict
what may happen to be useful, successful, or required in the future.
SRF

... it seems that there is a general agreement that "ontologies can
have individuals, but only in special cases".

Yes, but I don't believe that we have considered, analyzed, and
precisely specified all the kinds of special cases.

While that is true it seems that there are two issues here. Since you brought it up I'll consider it first. That is that in science we don't have to consider all the special cases immediately. We use constraints to specify the scope of any hypothesis that is given during the thesis or subsequently during discussion.

Second, I took a look at Gruber's paper that mentions T-Boxes (A Translation Approach to Portable Ontology Specifications). It is a 1993 paper about tools and their use for translating text and is more of a technology discussion. This is why I eschew discussions of tools.

 Neither wikipedia reference on T-Boxes or A-Boxes mentions the earliest reference and the wiki material on A-Box lists references that were subsequent to Gruber's. Both articles are cited for insufficient references. Perhaps we should go back and take a look at Whitehead and Russel's charges against the concept of Principia to understand the meta-reasoning in this area.

Yes, there is an ongoing discussion. However, we cannot make progress in these areas without a problem statement. The crux of the matter is that discussion addresses a linguistic problem and there are many ecological niches that yield different language requirements. Until we specify the problem to be solved, we cannot have a meaningful discussion about the road that gets us there.

(And I would like to note that there is an additional box in science that is useful in information science, the S-Box. I have seen it variously referred to as a Substitution-Box or a Shift-Box. Most of its use has been in crypto, and many of the papers with its use are classified, hence it gets little exposure.)

-John Bottoms

SRF

I would add that, besides being necessarily true, facts about specific
individuals can be present in an (domain) ontology only if they are
shared by the community.

This statement involves too many constraints such as "necessarily true",
"can be present", and "only if". And the term 'community' is much too
vague to be dependable. That statement might be useful as a guideline,
but not for a definition, standard, or requirement.
WA

I sympathize with your intuition, but it would take some work to
pin it down precisely.

Bill made this statement in response to Nicola, but it could be applied
to almost every general principle proposed in this thread or email list.
RD

uniqueness is a very important point to support mathematical analysis.

Yes. That is why functions are important in mathematics and logic.
Whenever you have a function, say x=f(y,...,z), it implies that there
is exactly one x for every combination of the arguments y,...,z.
HP

Don’t forget that identifiers for individuals are grounded in institutional
frames of reference with context and scope.

Yes. Every system of unique identifiers specifies a function from
a set of individuals in the universe of discourse to a set of symbols
that somebody chooses to call "identifiers". In practice, implementing
such a function and guaranteeing uniqueness is a non-trivial exercise.
AS

does anybody have Tbox with individuals?

The term 'Tbox' has never been precisely defined. But any geographical
information system must make provision for special individuals called
the earth, sun, and moon. It must also include many facts about them.
For example, the earth is an oblate spheroid, the moon revolves around
the earth, and the earth-moon system revolves around the sun.
Any ontology for geographical information systems that did not include
those individuals and many such facts about them wouldn't be useful.
JLC

I wonder if can we commit ourselves to an ontology that does not
distinguish between types and tokens.

Peirce used the type/token distinction in talking about signs. That is
an important topic. But when we're talking about ontology, it's better
to use metalevel terminology that is directly related to the logic.
To use Quine's criterion, the things that exist in your ontology are
those you can refer to with a quantified variable. Then types are
specified by monadic predicates. Some logics let you use variables
to refer to types (monadic predicates) -- that allows types of types.
HAK

Why has a specific formalism terminology (namely OWL's TBox, ABox, ...)
become standard? There have been generic names (Schema, Database,
Knowledge Base, ...) independent of any specific formalism - as it should be.

Yes. The field of ontology -- more generally, knowledge representation
-- has a large number of terms that are almost synonyms. In fact, every
notation introduces a set of terms that are not quite the same as the
terms used with other notations (even when they have the same spelling).
I prefer to start with a minimal vocabulary for logic: 'and', 'not',
'there exists', and either 'relation' or 'predicate' with the option
of zero or more arguments for each relation or predicate. Then
everything else can be defined in terms of those words.
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 
 
-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 201202221 / Virus Database: 2634/5453 - Release Date: 12/12/12

_________________________________________________________________
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