[Top] [All Lists]

Re: [ontolog-forum] Foundation ontology, CYC, and Mapping

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Rob Freeman <lists@xxxxxxxxxxxxxxxxxxx>
Date: Sun, 14 Feb 2010 20:09:56 +1300
Message-id: <7616afbc1002132309s604a5b80s150b7d9101d6a23c@xxxxxxxxxxxxxx>
John,    (01)

Can you tell me what text by me you're referring to here.    (02)

On Sun, Feb 14, 2010 at 6:30 AM, John F. Sowa <sowa@xxxxxxxxxxx> wrote:
> ...
> Note to RF:  Yes, the patterns of words in NL text impose strong
> constraints on the meanings of the words.  That is important for
> NLP, but more explicit spec's are important for computer software.    (03)

If it is my plea that we stop disputing names for things, I think that
applies even more to computer software.    (04)

There is even less need to use fixed labels in computer software.
Certainly not at the application level.    (05)

There is some excuse for using names/labels in human conversation,
because humans have bandwidth problems. I can't hand you a whole set
of events, or generalizations about events, easily as a human, so I
use a word in the hopes it will evoke roughly the same set of events,
and generalizations (mappings and relations?) about those events, in
you.    (06)

The match is often not very good, unless two people have narrowly
similar experience and habits, but there is no alternative.    (07)

But computers have no such problems. They don't need to work with
labels, they can work directly with the sets of events or
generalizations about those events, the labels refer to.    (08)

It might be a bit of a challenge initially to write a program in terms
of such sets of events. But that is a problem of human conception and
expectation again, not anything intrinsic to computation. Computers
are quite happy for a computation to have no decidable axiomatic
formulation.    (09)

Equally they can make whatever generalizations you like, and on
request likely work back to a decidable axiomatic formulation relevant
to one or other special case, if one exists.    (010)

I suppose you could limit your computer applications to those with
decidable axiomatizations. Maybe that's a good description of what we
do today. But then they wouldn't model meaning very well. This would
prevent useful applications in any domain where you seek to get near
cognitive performance: perception, judgment, image recognition, not
just language.    (011)

But I am just guessing what point you are arguing to. If you can tell
me what statement by me you are referring to I might be able to say
more.    (012)

-Rob    (013)

P.S. I like your argument against meaning primitives in axiomatic
terms. It makes the point very strongly.    (014)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (015)

<Prev in Thread] Current Thread [Next in Thread>