Rick, with respect to Furman and
Gallo, I understand the basis of your criticism but have to disagree with
your conclusion that by equating PCT with Miller et al’s TOTE mechanism,
Furman and Gallo “got PCT completely wrong.” True, Miller
et al. describe TOTE as a sequence whereas PCT has all parts of the loop
simultaneously active. But TOTE is simply a type of control system
found at the HPCT program level, so to say that TOTE gets PCT
“completely wrong.” is a bit of an
overstatement.
Both types of control system
include a reference level or goal state. In both, this reference
level is compared to a current perception. In both, any difference
between the two results in further action being taken. In both, the
effect of this action on the current perception is to reduce the error
between perception and the reference.
That is why I stated above that TOTES fit into PCT’s program level: the
next action in the sequence cannot begin until the previous one makes
some “if-then” statement true (e.g., if cup at lip then begin
sipping).
[From Bill Powers (2008.01.02.0959)]
Happy New Year, All!
Bruce Abbott (2008.01.02.1140 EST)]
I somewhat agree with you, but the TOTE unit is really not any kind of
system model – it’s a description of what the same system is doing at
different times. The key is the “E” in the name: this
“unit of behavior” is modeled after a computer subroutine, from
which it is possible to “exit.”
I was in communication with Pribram for a year or so before 1960, when
both “Plans and the organization of behavior” and “A
general feedback theory of human behavior” were published. Nothing
much came of the correspondence, but Pribram was aware of the work of
Powers, Clark, and MacFarland when “Plans” was published. I
don’t believe we were cited.
In 1960, “sequence” was the fourth level of control, just below
“relationships.” That level finally turned into
“events” sometime in the 1990s, with sequence moving up to
level 8. Actually, the TOTE unit fits best at current level 9, the
program level, which involves tests and choice points. Sequences would
describe the fixed lists of processes that take place between tests –
Pribram’s “operations.”
I agree with Rick that generally, attempts to link PCT with other
theories have been pretty unsuccessful. What usually happens is that the
other person is looking for validation of his own theory. The person
tries to find things in PCT that agree with his theory – which usually
results in some pretty gross distortions of PCT like the ones in that
book excerpt about NLP, which I’m sure everyone here noticed. It’s pretty
rare to find another theory linked to PCT in a way that shows any modest
degree of understanding of PCT.
I really think Rick is right in saying that the only meaningful
comparisons are between a theory and the relevant data. Theories should
be compared not on the basis of how well the terms of one theory
translate into terms of the other theory, but on the basis of how well
each theory explains and predicts observations.
Best,
Bill P.