On May 6, 2006, at 11:40 PM, John Scanlon wrote:
A typical human being's knowledge of the letter "A" involves
recognition of graphical representations of the symbol, memories of
its sound when spoken, procedural or muscle memory of how to speak
and write it, and memories of where it is commonly found in its
linguistic context. A system should be capable of recognizing
symbols visually or auditorially (and possibly of generating them
through motor outputs) before it should be expected to comprehend
them.
I would point out that none of these symbols (auditory, visual, etc)
is equivalent, except in arbitrary and narrow contexts. Your
specific brain contains a context in which they are conditionally
equivalent, but the letter "A" in all its forms has no universal and
objective information theoretic nature. It seems a bit loose to
define "comprehension" in terms of an arbitrary collection of soft
correlated pattern equivalences in some narrow context. What,
exactly, is there to comprehend about the letter "A"? It is a
pattern, nothing more, though it is sometimes correlated with other
things in many other contexts. If only having a specific set of
correlations mattered, Cyc would be real AI.
Cheers,
J. Andrew Rogers
-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
