Christopher, (01)
I am responding to this only because it was directed to me. (02)
you wrote: (03)
> I support everything you said here (below). But - in the light
> (or darkness...) of a long-running recent discussion on this
> forum - are you carefully steering clear of the word "context" on
> purpose? :-) (04)
Well, I actually used the word, but I'm carefully steering clear of
trying to define it. (05)
> Is this not a prime example of where contexts - and conversions
> or translations or mappings (often lossy) between them - are used
> by everybody with no problem? It's as in colloquially-familiar
> and meaningful statements such as "The precise meaning of the
> time statements is dependent on their context." (06)
I disagree with "used by everybody with no problem". The context is
implicitly used by everybody *in the in-crowd* with no problem. The
problems arise when the statements are used by *anyone else*. And the
particular problem that concerns me is the conclusions that will be
drawn by software agents, and automated reasoners in particular. (07)
> Can we really do without some formal recognition and
> representation of 'context', all in a context-dependent way, of
> course? And surely, there is no need to eschew the word! (08)
This is exactly the point, or rather the two points. (09)
(1) You can formalize statements without "formal context" only as long
as the reasoner is implicitly operating in that context (for all of the
assertions it is using) and everyone who uses the conclusions of the
reasoner understands that. If we are going to build public ontologies
for arbitrary use, the reasoners using them will, in general, not be
operating in a predictable context. So we have to be very careful in
our statements, or very clear that we don't expect "arbitrary use" of
the ontology. (I personally think we always need to do both.) (010)
(2) "context" is context-dependent. The literal interpretation is:
whatever else I need to assume in order to understand what you meant.
And formally that definition means "some arbitrary hypothesis base".
To formalize a "context", you need to formally state all the assertions
and antecedents that you are assuming. So there is nothing special
about "context" in the large -- it is just another "body of shared
meaning" (as the SBVR folks say). (011)
AI efforts to capture "implicit knowledge", to allow reasoning agents to
communicate effectively with people, have been appearing on-and-off for
30 years. But much of that work is about enabling the reasoner to make
the logical links between assertions that the human never bothers to
state: "We drove to the concert, but we missed the warm-up band because
we couldn't find a parking space." But much of "context in the large"
is the knowledge related to common practice, just as it is in that
statement. (012)
-Ed (013)
P.S. I intentionally distinguish between "context in the large", which I
think is what Christopher is talking about, and the "specific context"
of a vocabulary term. "In mathematics, a 'field' is ..." or "when said
of an electrical lead, 'hot' means ..." Mechanisms for expressing
"specific context" for the interpretation of a symbol are
well-understood. But to apply them "in the large", you would have to do
it on an instance-by-instance basis. (014)
"You may believe you understand what you think you heard, but
what you do not realise is that what I said is not what I meant."
-- R. Alexander (Sandy) Tyndale-Biscoe (015)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 (016)
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority." (017)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (018)
|