[Top] [All Lists]

Re: [ontolog-forum] Past, Present, and Future of Ontology

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
Cc: Cyclify Austin <cyclify-austin@xxxxxxxxxxxxxxx>, KR-language <KR-language@xxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Fri, 29 May 2009 00:10:42 -0400
Message-id: <4A1F6042.60406@xxxxxxxxxxx>
Dear Matthew and Dick,    (01)

MW> I agree that context is used for what you cannot be bothered to
 > make explicit, and the answer is simple -- be bothered to make it
 > explicit, then you have no need for context.    (02)

To a certain extent, that is true.  Many different theories of context
have been developed, and all of them share one common feature:  they
use metalanguage to say something along the following lines:    (03)

  1. p is the proposition or sentence at the focus of attention.    (04)

  2. C is a conjunction of propositions called the context.    (05)

  3. p is related to C by certain relations r1, r2, ...    (06)

Most languages that deal with context state that kind of information,
and they also specify rules of inference for reasoning about that
information.    (07)

MW> I think "context" is way over in the woolly corner and needs
 > some disambiguation, deconfliction and deconstruction (and
 > probably some other words beginning with "d") before it can live
 > in anyone's ontology.    (08)

What makes context so woolly is the very difficult problem of
extracting or distinguishing the proposition p, the context C,
and the relations from some natural language text or dialog.    (09)

In addition to the basic problems of natural language processing,
which are by no means simple, you must, for each sentence, determine
what is the proposition p, how much of the surrounding verbiage and
the nonlinguistic environment should be considered the context C,
and what are the relations between p and C.  After you've done
all that analysis, then and only then can you map the results to
the neat formalism called a context language.    (010)

I agree that the process is uncertain and woolly.    (011)

RHM> I do appreciate your sporadic attempts to encourage me.    (012)

Thank you.  When you first contacted me, you seemed to be an intelligent
person who had been doing some hard work.  I am always willing to help
bright people who honestly want to learn.    (013)

But when you claimed that the work you did in the past 5 years with
no prior study in the field was the equivalent of what all the people
working on Cyc had done in the past 25 years, that was so ridiculous
that it immediately classified you as a crank.    (014)

Even Newton said that he was standing on the shoulders of giants.
Anybody who claims to start from scratch and reinvent everything
is just a crank.    (015)

And by the way, I am not saying that as an academic.  I also spent
30 years working for IBM, at which I was expected to do something
that might contribute to the bottom line.  And I am now president
of a small company (VivoMind Intelligence, Inc) which is actually
producing software that people are willing to pay money for.    (016)

RHM> Syntactically, context is represented in the mKR language as
 >    at space = s, time = t, view = v { sentence; };
 > where s, t, v is the context of the sentence.  v is the name of
 > the context, which is implemented as a complex data structure
 > in the mKE program.  s, t name the space-time coordinates which
 > are applicable when sentence involves a "physical" situation
 > such as an action or interaction.    (017)

This kind of representation falls right in the middle of the road
of the kinds of things that many people have been doing.  It has
the three parts:  proposition p, context C, and relations space,
time, and view.  But it lacks the rules of inference that specify
what kinds of reasoning can be performed on that notation.  Your
statement about the "complex data structure" v is also rather
woolly and requires a more precise specification.    (018)

What Pat, Chris, and I have found very irritating is that you keep
making grandiose claims about having a context language that solves
all the world's problems.  But in fact, it's similar to what other
people have been doing for the past 30 years or more.  Unlike those
other systems, your specification lacks rules of inference.    (019)

RHM> I think in human terms.  Start with a "tabula rasa" brain --
 > where the only propositions are those which correspond to the
 > structure of the human brain, which recognizes "existent", "entity",
 > "part", "attribute", etc.  Then progressively, fill in the context
 > with all the propositions which represent the human's sensory
 > experiences of the external world, and his thinking about those
 > experiences.    (020)

In Matthew's words, that is truly woolly.  Aristotle was also thinking
in human terms, but he was a very careful analyst.  In his books on
interpretation and the psyche, he went into great detail about all
those issues.  In the past two millennia, the greatest philosophers,
psychologists, linguists, and neurophysiologists have gone considerably
further, yet many unsolved problems still remain.    (021)

You can do good work in areas of your expertise without having studied
all those writings.  But when you claim that your woolly paragraph
is somehow superior to those two millennia of research, you make
yourself sound like the crankiest of cranks.    (022)

Suggestion:  If you stick to the syntactic and semantic details of
specifying your notation and defining what it means in computable
terms, then there's something we can usefully discuss.    (023)

But when you make grandiose claims that your thoughts are somehow
superior to all those publications you have never read, then we
just dismiss you as a crank.    (024)

John    (025)

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    (026)

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