X-Message-Number: 4047
Date: Sun, 19 Mar 1995 20:17:24 -0800
From: John K Clark <>
Subject: SCI.CRYONICS Hofstadter

-----BEGIN PGP SIGNED MESSAGE-----

In #4044  on  Sun, 19 Mar Bruce Zimov <> Wrote:
   
	     >Since GEB was written in 1979, it predated the work of
	     >Rumelhart and  McClelland in PDP which explained alot of what   
	     >Hofstadter, coming from  the traditional Minsky/Papert AI 
	     >perspective, was overtaxing the concept  of 'symbol' to do.
	      
You've badly misjudged Hofstadter. Fist of all, he's one on the
leading critics of the traditional "expert system" approach to
AI, maintaining that it's too rigid and fragile to work, except
in very narrowly defined fields, and second I've never heard him
say anything about Rumelhart and McClelland's work that was not
full of praise.
	      
	     >The last 10 years of connectionist research in vision  and
	     >other areas stands as a convincing argument that Hofstadter's   
	     >symbols  are a historical curiosity at this point.
	    
For the last 10 years Hofstadter has been writing experimental
programs ( Jumbo, Numbo, Copycat ) that employ a philosophy
similar to that of Rumelhart and  McClelland, they call it
subsymbols, Hofstadter calls it active symbols, but both are
highly distributed and use vectors in multidimensional space
that are not fixed and so can change position as needed. If the two
concepts are not identical they are certainly first cousins. 
	    
	     >It is interesting that in the  quotes above, Hofstadter
	     >regards consciousness  as 'liftable' but not to any abstract
	     >nonphysical level.
				 
Well I should hope not!!! If Hofstadter started talking about
religious garbage I wouldn't waste my time with him.  
			 
	     >No matter who is correct on the symbolic/ sub-symbolic issue
	     >for intelligence, one can see that the 'liftability' of         
	     >consciousness is a separate issue
	      
I agree completely. Exactly how the brain uses symbols is a
fascinating philosophical problem, but not terribly relevant
when discussing the possibility of uploading. The brain has some
PHYSICAL mechanism for dealing with symbols and whatever it is
it can be duplicated. We don't even need to figure out what it
is to do an upload, just blindly copy from nature.
	      

			       John K Clark              

-----BEGIN PGP SIGNATURE-----
Version: 2.6.i

iQCzAgUBL2z5uH03wfSpid95AQGLqwTuMrrmzubjygfqZ26ovjPlVsmC2dlmKtAq
9/YxZJe7do70Eq53SLsFJjiP1sQ7iFo3HTp+pK7y+BnjM0V/nol+Nerk51dQVccb
nbWe2p0FAsGNvhmVD/50kqBkxFiFm44tMbsfW87ZPCQ27dg+b10CtOq8lf6L8eSe
zhVSlOkE/0ZEe5T+5fODfruIEm+BdeX2kYvl5eQCpbuCM06K96E=
=F2sK
-----END PGP SIGNATURE-----

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=4047