[Martin Taylor 921216 11:30]
(Bill Powers 921215.1530)
Bill, I think my conscious impression of what one imagines during planning
is very like yours, and I can imagine both your and Mary's version of
the golf ball in the Coke bottle. But none of it addresses the point
that I wanted to raise. Perhaps I can try another approach.
All the perceptions involved in the conscious imagination are, by hypothesis,
obtained from perceptual signals in the same ECSs that would be involved
in actually executing the plan. When the plan is executed in the real
world, these perceptions are derived from sensory signals that reflect
the reactions of the world to the actions evoked by the outputs of the
ECSs at the various levels involved. When the plan is executed in
imagination, the perceptions are derived from sensory signals that
originate in some other place.
Whenever this question has been discussed, on CSG-L or in BCP, that
"other place" has been either a memory store or a looped-back output
signal that normally contributes to the reference signal of a lower
ECS. In the first case, the "sensory" signal is one that has occurred
before in the real world; in the second, it is what would be sensed if
the reference to the lower ECS were to be satisfied by the actions of
that ECS controlling its perceptual signal. But in the real world, some
imposed constraints may prevent that lower ECS from bringing its percept
to its reference level. In the planning situation, these constraints
have never yet been encountered in the real world. They are imagined
as a consequence of something else, such as verbal instructions.
In the versions of imagination that I have seen discussed, I see no place
for simulations of a real world never experienced. It begins to feel as if
some construct other than an ECS is necessary so that a variety of
imaginary-world impedances can be placed in the imagination loop, to
match the impedances that might be found in the real world for which the
plan is being made.
It is important for the survival of the organism that the possible variety
of impedances be insertable in the imagination loop, especially in situations
where there is danger (implicit in some of the possible impedances that
might be encountered with different action choices).
I still feel that the necessary structures may exist within the hierarchy,
but I can't find them. Pushing the question to lower levels only moves it;
the problem stays the same.
Martin