[From Bill Powers (2010.02.25.1123 MST)]
From Richard Kennaway (2010.02.25.1036 GMT) –
[From Bruce Gregory
(2010.02.24.2008 UT)]
BG: You seem to be ignoring the possibility that inhibitory connections
can be strengthened. If you are at all interested in learning something
about this topic you might look at Neural Networks and Animal Behavior
by Magnus Enquist and Stefano Ghirlandia (Princeton,
2005).\
JRK: What an extraordinary book
that is. I’ve only looked at the pages available on Google Books
and Amazon, but I think that’s enough.
…
Models? Here’s a model
from chapter 1:
r =
m(x)
x is the animal’s state and r is its response – its behaviour. The
author calls m the behaviour map, and says it is common to all theories
of behaviour.
…
The first few pages of Chapter 1
make it clear that the book is solidly grounded in the assumptions of
stimulus-response operation and behaviour controlled by
perception.
BP: That’s the impression I’m getting so far from neuroscience in
general. It really does look as if S-R has simply migrated from
psychology into neuroscience – though in truth, it probably began with
neurology (and biology) and has simply stayed there. Watson’s
initial formulation of behaviorism is said to have originated in biology.
This makes me wonder about the number of departments of psychology I have
seen which are now called departments of psychology and neuroscience, or
neuropsychology, or some other term with a “neuro” prefix used
somewhere. The “neuro” part looks like a signal that says
“Stimulus-reponse spoken here.” But I don’t really know if
that’s fair. My sample is very small – though what psychology does get
into Science and Nature is pretty solidly on the side of
SR, and that indicates a pretty wide penetration, or chronic infection,
by SR ideas.
If you look at the basic control-system diagram we use in PCT, you can
see that it’s possible to trace signals from sensory receptors to more
central structures and back out again to the muscles. If that’s all you
consider, the only possible model is the SR model. A neurologist using
this concept as a guide for exploring the nervous system can obviously
trace cause and effect, progressing synapse by synapse, from input to
output. Why look for anything else?
The problem is that this sort of experimental investigation involves
applying inputs under control by the experimenter, which prevents any
effects by the organism’s actions on the stimuli while they’re being
applied. This guarantees that no concurrent feedback can occur. Because
of that, it’s impossible to see any sign that the actions are organized
to have specific effects on the inputs. And this means that it’s
impossible to see that these feedback effects tend to maintain input
variables in specific states that we would call reference levels, which
means that nobody will ask how those reference levels are specified.
Nobody will discover the existence of reference signals in the brain.
That means that nobody will discover the hierarchy of control. All of
which is what happened.
There are some current opportunities for me to sneak PCT into some
departments that are joined with neuroscience. It seems clear that the
main thing to do, to make any sort of worthy contribution, is to get
people to see and understand the feedback loop, the control system idea.
It’s probably unnecessary to do much more than that, other than thinking
of more demonstrations and experiments to make the understanding more
solid. Once neuroscientists understand how feedback control works (I
haven’t yet met more than a couple who do), they will see that there are
a lot of new phenomena to explain, and a lot more kinds of neural
functions to account for. Suddenly a lot of new pathways will be found in
the nervous system, not because they suddenly grew, but because the
investigators were not equipped to recognize what they were looking
at.
It’s fortunate that my new book is subtitled “the fact of
control.” I wasn’t thinking quite this way while I was writing it; I
just wanted to get the demonstrations out there for people to look at.
But now it’s clear that the main thing we have to do to bring PCT into
mainstream science is to convince neuroscientists and the rest of them
that control is a fact of nature that has to be taken into account. They
will know what to do about that fact once they decide to accept it, and
once they realize that a lot of what they know about feedback ain’t so.
If we can just accomplish that much, I think the revolution will become
self-sustaining.
Best,
Bill P.
What I have learned about modern neuroscience so far shows that it has
made a lot of progress in exploring neurological phenomena. It’s obvious
that I still have a lot to learn about the state of this field today, as
opposed to what I could find out before publishing B:CP. But it’s also
obvious that the weakest part of neuroscience is in its idea of what
behavior is and how it works. What we have now is a picture of very
sophisticated experimental methods being applied to an outdated and
inadequate concept of the behavioral phenomena to which neuroscientsts
are trying to link the neurological and biochemical findings.