[From Rick Marken (930925.2000)]
Bill Powers (930925.0800 MDT) to Avery Andrews (930925.1215) --
Your example of fixing the lawnmower brings up a lot of
fascinating problems.
Hearty agreement on that.
I'm on this tack because of having been reading BBS
Oh my god. Next it'll be AI Magazine.
I'm amazed at how some people try to make words
and symbolic operations accomplish everything.
One reason for this (I think) is that this is what digital
computers are good at. As soon as somone figured out that
100101101 could be a symbol as well as a quantity, word
(or symbol) based simulation was on its way. This would
never have happened if the computer of choice became the
analog rather than the digital computer. Of course, if the
analog computer were the only kind we had, psychologists
would have all been PCTer who might now be having trouble
accepting the fact that analog variables can be symbols.
Anyway, one nice thing about the word based approach to
simulation is that it obviates the need to be able to
do what PCT models will almost certainly not be able to
do for some time -- model the perceptual functions
of higher level variables.The perceptual functions that
have been built produce pretty low level perceptions (from
the point of view of the hierarchy) -- intensities,
sensations, configurations, transitions, possibly
events, categories, but certainly not programs,
principles or system concepts.
I think there MIGHT be a way to do some modelling of
"higher order" control using the "word based" approach
of the BBS crowd -- but with a PCT flavor (maybe this
is what Michael Fehling was suggesting). Perhaps
there is a way to use the words as the "addresses"
of "perceptual variables" that are analog in the
PCT sense. The goal would be to finesse, somehow, the
problem of trying to build perceptual functions for
the variables represented by the words -- but have
the variables functions "appropriately" in a perceptual
control hierarchy.
I don't know how to do this; I'm just free associating
like you were (your associations were better than mine
but I think we both have to pay Avery the same amount
for the treatments -- so I'm getting a deal).
Maybe the "word based" approach is impossible for PCT
modelling -- but I think it's worth a little thought.
Whaddaya think?
Best
Rick