RE: [humanmarkup-comment] Referents

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

http://www.ecs.soton.ac.uk/~harnad/Hypermail/Foundations.Cognitive.Science2001/0158.html

Worth reading.  Discusses the symbol grounding problem. 
A sign system has this too except that we are looking 
at an adaptive system that does react to the environment 
in order to ground the symbols.  That is why the 
proximity and sensory models are revealing for our 
work.  Think of it was working backwards from the 
referents to the signs, then in reverse once the 
signs are "learned" or "taught".

I think he is wrong about emotions.  I believe that 
is the blind spot of both classical and hybrid AI. 
If the emotional response is simply a layer above 
the physical reflexive response, we most certainly 
can model these in a sign system.  The simplifying 
assumption is that again, can submarines swim: you 
don't have to know the answer to use the sub.  We 
don't have to know how computers "feel".  We only 
have to enable them to choose signs to tell us.

len