To clarify, the approach I describe is open to use of formal
languages such as predicate calculus and conceptual graph
structures, and to use of other notations and representations
when they can more clearly or concisely express information than
natural language. It could also support local jargon, acronyms,
etc. So while the focus of my thesis is on what can be done with
a knowledge representation language based very largely on the
syntax of a natural language, I would agree with your point that
"unnatural language" is also part of the challenge.
From:
deddy@xxxxxxxxxxxxx
Date: Wed, 23 Apr 2014 10:06:00 -0400
To:
ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Toward Human-Level Artificial
Intelligence
John -
On Apr 23, 2014, at 9:47 AM, John F Sowa wrote:
The intro summarizes the
arguments for the
controversial claim that human-level AI is possible
with a knowledge
representation based on natural language.
And what about the issue that a goodly slice of human
communications is decidedly NOT natural language?
Acronyms, local jargon, industry slang, puns, etc.
I'm specifically thinking of your VivoMind effort where
you well know the closer one comes to the "coal face" (the
working system) the weaker the correlation to human
language as found in commonly accessible
dictionaries/glossaries. Wasn't something like "computer"
used as a mythical "person" to charge billable time to?
All I'm arguing is that unnatural language is part of
the challenge, along with natural language (which is
challenging enough). To ignore unnatural language is
artificially simplifying the problem.
Example:
- the CEO's shareholder letter is natural language
- the specifications for a software product are likely a
mix of natural language & localisms
- the actual code is likely to be very heavily unnatural
language.
____________________________
David Eddy
Babson Park, MA