I agree with Adam that this question seems to be an odd topic for a workshop.
For me the practical answer is very simple, and to go beyond that
doesn't seem to have much purpose. It may be true that a lot
of people who are just starting to create ontologies don't
really grasp how one specifies conceptual content using
axioms to define predicates -- but that would be a topic
for a tutorial, not a workshop.
It is hard to avoid the question completely because in discussions
of knowledge representation generally and ontology in particular,
the term "meaning" is used a lot, expected to be interpreted in
its intuitive sense, but it is natural for readers to wonder
if there is a specific definition.
When I say "meaning" I mean this: a text carries information
that allows an intelligent reader (human or machine) to make
alternative **choices** based on the structure of text -- words,
images, and and format. The document content symbolizes conceptual
structures in the mind of the agent that created the text. The more
information a text has, the more choices are enabled, and the more
(quantitative) meaning the text has. The information content depends
on the intention of the agent constructing the text, since it might not
be interpretable by any other agent (if, e.g., encoded in a private
code). The "meaning" itself is ***the set of conceptual structures in the
mind of the creating agent that are symbolized by
the text and its format***. When the agent reading the text
determines that the symbols refer to a specific set of conceptual
structures, and these are the same as the conceptual structures
in the mind of the creating agent, the reading agent has
correctly interpreted the "meaning". An agent can determine
the meaning of some parts of a text and not others.
Determining the meaning (conceptual structures) of any part
of the text enables the reading agent to make choices based on
that information. An ontology provides a formal set of symbols
for conceptual structures which, if used by both the
agent creating a text and the agent reading a text, allows
the reading agent to reproduce the conceptual structures
which were in the mind of the creating agent and were
encoded in those symbols. The predicates get meaning by their
definitional implications using axioms, which provide symbols
whose relation to fundamental conceptual structures is widely
understood.
To contrast this, we know that file transfer protocols can
take a file on one computer and create a copy on another computer.
When the protocol creates an exact copy and the action of the
receiving computer does not depend on any element of content
of the file, the computer does not use any element of
"meaning" encoded in the file.
But as soon as the computer reading a file can make some
**choice of actions** depending on the content of the file,
it has interpreted the file in some way and can be said to
have used some element of "meaning" in the file. This can
be very trivial, as when a transfer protocol converts
cr-lf line ends to lf or vice-versa. In such a case,
the computer has interpreted one part of the format, which
is (or could be) part of the meaning of a file. So there
is a continuum of meaning, from the very trivial to the
very complex.
At a slightly more complex level, a computer might receive
a file labeled "invoice" and because of that label, search for
the fields labeled "amount due" and "pay to", and then
print a check in the amount due to the payee named.
The computer has then interpreted some elements of
meaning encoded in the text. If the computer also scans the
line item fields in the invoice to determine whether
the items received are those the invoice says were
sent, it is interpreting other elements of meaning in
the invoice. (01)
This is a very practical use of the term -- the meaning of a
text is the set of conceptual structures that the creating agent
intends it to symbolize, provided that the agent has encoded that
meaning (the conceptual structures) correctly. (02)
I don't think we need a conference to debate these issues.
If it gets much more complicated than the above, then, as Adam
says, the discussion would get too philosophical and distract
from the practical tasks. (03)
Pat (04)
====================================================== (05)
Adam Pease wrote:
> Folks,
> I'm kind of surprised at the workshop text. This is an issue that has
> been debated at length for decades in computer science, and over a
> century if one expands to semiotics. Some researchers like Rodney
> Brooks take a strong situatedness view that symbols must be ultimately
> grounded in the physical world to have meaning. Others take a "brain in
> a vat" view that meaning, and by extension, intelligence, need not be
> situated. Ultimately I view this as an interesting but unresolvable
> philosophical debate, without much impact on practical ontology
> building, and on which I wonder whether much additional light might be
> shed. Building systems, whether situated or not, is unlikely to address
> this issue. In fact, I'd venture that this is the sort of topic which
> can have a substantial negative impact by distracting from the practical
> tasks of building and using ontologies.
> I'll go back to my grumpy guy cave now :-)
>
> Adam
>
>
> At 07:21 PM 6/3/2004 -0400, Internet Business Logic wrote:
>
>> Hi All --
>>
>> In case you have not already seen it, there is a very interesting call
>> for workshop papers [1].
>> The workshop theme begins:
>>
>> "What do ontologies, as used in the semantic web and elsewhere, have
>> to do with meaning? In particular, where do their predicates get their
>> meanings? Semantics, no matter what formalisms are applied to it, is
>> ultimately a cognitive phenomenon: it refers to the meaning that
>> symbols have for human beings. ...."
>>
>> This is a question that is also raised at a nuts-and-bolts level in
>> [2], where a technical, implemented answer is attempted.
>>
>> Any thoughts or pointers about [1] and/or [2] ? Peter -- is this a
>> suitable topic for an extended discussion on this forum?
>>
>> Thanks in advance, --
>> Adrian
>>
>> [1] http://fois2004.di.unito.it/workshops.html
>>
>> [2] "Semantic Web Presentation" at www.reengineeringllc.com
>>
>>
>>
>> --
>>
>> Internet Business Logic -- online at www.reengineeringllc.com
>>
>> Reengineering LLC, PO Box 1412, Bristol, CT 06011-1412, USA
>>
>> Phone 860 583 9677 Mobile 860 830 2085 Fax 860 314 1029
>>
>>
>>
>>
>> _________________________________________________________________
>> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
>> Subscribe/Unsubscribe/Config:
>> http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
>> Shared Files: http://ontolog.cim3.net/file/
>> Community Wiki: http://ontolog.cim3.net/wiki/ To Post:
>> mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Subscribe/Unsubscribe/Config:
> http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/ To Post:
> mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
> (06)
--
=============================================
Patrick Cassidy (07)
MICRA, Inc. || (908) 561-3416
735 Belvidere Ave. || (908) 668-5252 (if no answer above)
Plainfield, NJ 07062-2054 (08)
internet: cassidy@xxxxxxxxx
============================================= (09)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (010)
|