[From John Anderson (950629.0630 EDT)]
I've recently finished reading B:CP, and I'm reading _Intro to Modern
Psychology_ now. After IMP, I have LCS I and LCS II to read. So I'm
learning the nuts and bolts of PCT. I have a few comments and replies
to postings from the last month or so. Sorry to have let so much time
elapse.
···
---------------------------------------------------------------------------
Bill Powers (950618.1430 MDT) --
What has me climbing the wall is the use of the term "selection by
consequences" when to me it is obvious that nothing of the sort is going
on......Am I the only one in the world who sees statements
like this as invoking magic? Here is a consequence; here is a contiguous
relationship. Then a miracle occurs, and reponses are modified or
behavior is selected. Doesn't anyone else see something vital missing
from these statements?
I do. What's missing is biology, to which both you (950620.1235 MDT)
and Dennis Delprato (950620) have alluded. Consider the following
hypothetical physiological mechanism, which seems to me to be
compatible with both PCT and reinforcement, at least as I presently
understand them.
Suppose that the results of an action, like choosing a certain golf
club, reduce the error between some desired (reference) perception,
like getting the ball in the hole, and the actual perception. Now if
some process strengthens the synaptic connections in the control
systems whose activity produced the action that led to those "good"
results, then when those control systems become active again, a
similar neural response will be likely to occur. If the global set of
active control systems in which it occurs is similar to the one
preceding the first "good" result, which I think is a reasonable
expectation given a similar situation, then the overall action of the
organism (the golfer) will probably be similar, and hopefully (for the
golfer) will lead to a similar result. This is a sort of reverse
reorganization: when the error becomes small, the current organization
-- the pattern of synaptic connectivity and synaptic weights -- is
made stronger. Conventional PCT reorganization occurs when error gets
large: the current organization is weakened, and an alternative one is
strengthened, so that in a similar situation, the "wrong" action does
_not_ occur. Either way, the actual result is compared with a desired
result to determine if it is "good" or "bad", and the neural machinery
is modified accordingly. This hypothetical mechanism is admittedly
low in detail and needs refinement, but you get the general idea. It
seems compatible with PCT, and, if you ignore the biology, also seems
consistent with selection by consequences, at least as it has been
described on CSG-L.
---------------------------------------------------------------------------
Bill Powers (950519.1325 MDT) --
John Anderson (950519.1045) --
Here's a blurb about Walter Freeman's new book that sounds quite
interesting.Since you've read the book, I'd like to ask a few questions.
Sorry, I have not read the book, but I can answer one of your questions:
Recent studies of brain chemistry in animals during reproductive
behavior have revealed the operations of neurohumoral mechanisms,
which have evolved to promote success in mammalian reproduction and
the care of altricial young."Altricial" is not in my dictionary. What does it mean?
From the Hypertext Webster Interface at
http://c.gp.cs.cmu.edu:5103/prog/webster:
al. tri. cial \al-' trish-*l\ aj [L altric-, altrix fem. of altor one
who nourishes, fr. al] tus pp. of alere to nourish - more at OLD :
having the young hatched in a very immature and helpless condition so
as to require care for some time
From you:
Has anyone tried to construct a PCT model of Parkinson's disease?
I'm thinking of proposing to do this in the grant I'm working on.
I'd appreciate any info.What is it about Parkinson's disease that you want to model? The
oscillations are fairly simple to reproduce; just set the parameters of
the model so it begins to oscillate. This, of course, only hints that
the symptoms result from changes in control parameters; it doesn't tell
us anything about what is causing those changes.
Part of the grant proposal will be to relate what is currently known about
human neuroanatomy to the PCT control hierarchy; presumably (PCTers hope)
a PCT-like hierarchy will be there in some form. As we learn more about
the anatomical hierarchy, we can begin to build a computational model of
it (you folks have already begun this, right?). The most reasonable
strategy, I think, is to work in from the periphery, a process begun in
B:CP. If and when we get as far as the basal ganglia, where at least some
of the problems leading to Parkinson's occur, we can ask how well the
computer model matches the behavior of normals and Parkinson's patients in
behavioral tasks designed from the PCT perspective (suggestions for these
will be welcomed). My neurologist collaborator, Vinod Deshmukh, has a
large population of Parkinson's patients from which to obtain
participants. Data from these behavioral experiments will be used to
refine model parameters; it may also provide novel insights into a
common neurological disorder.
---------------------------------------------------------------------------
From time to time somebody refers to PCT papers having been rejected for
publication in various journals. What reasons do the reviewers cite for
rejecting the papers?
---------------------------------------------------------------------------
John