[From Bill Powers (951008.0815 MDT)]
Bruce Abbott (951007.0935 EST) --
Dammit, why do you always have to be so @#$%&*)! reasonable?
I have a million
reasons why it is not a good way to find out about behavior, but every
one of them will be met by a counterargument that will satisfy you.
I wonder whether you are really talking about me here or yourself.
As I was not given the chance to provide such counterarguments, I
can only conclude that you were able to think of them yourself.
You say you have a million reasons and yet you fail to provide even
one. Perhaps you believe that your reasons are just not convincing
enough to present.
OK, let's skip the statistical quibbles and go right to the biggie.
Why do you think that there is any specific effect of MSH on loud
persistent peeping? Simply because when you inject MSH, loud peeping
persists under conditions where otherwise it would diminish or cease?
This is certainly an empirical fact, but what does it tell us that is of
any use in understanding the brain or behavior? MSH, as you point out,
stands for "Melatonin Stimulating Hormone." What MSH should have done
for these chicks was to make them appear tanned under their feathers --
did anyone look? Or, according to some recent studies, MSH should have
enabled these chicks to sleep better at night (that's why you can now
buy melatonin pills at health food stores). I don't know of any studies
with humans of the kind you describe, but apparently MSH ought to have
an effect on human distress vocalizations under at least some
conditions. Do you really expect that this would happen?
Brain chemicals are part of the normal means of operation of the neural
system, just as traces of boron are part of the normal operation of a
transistor. When you artificially inject MSH, you are doing something
parallel to injecting a little extra boron into some of the transistors
that make up a radio. Naturally, this is going to have an effect on the
operation of the system. In the chick with MSH in its ventricles, you
get a change in the relationship between visual perception of other
chicks and loud persistent peeping (and who knows how many other effects
that you weren't looking for?). In the radio with boron added to some of
its transistors, you might get a decrease in the ability to separate one
station from an adjacent one, or an instability in the automatic volume
control, or a loss of bass response from the speakers, or distortion at
high volume settings, or an overall improvement in sound quality. If you
always injected the boron in the same place, you would probably get a
repeatable effect with the same make of radio.
Suppose you found that injecting boron at a certain position on the
integrated-circuit chips of a number of Hitachi radios resulted in a
loss of bass response in all of them. Would you then conclude that
"boron injections suppress bass response in Hitachi radios?" If you were
to publish such a finding, you would have electronics engineers rolling
on the floor and gasping for breath in hilarity. If one of them could
recover enough to speak, she would say, "Well of course, you idiot.
ANYTHING that reduces the gain of a transistor in the bass control
circuit will reduce the bass response. You could drip acid on a resistor
and get the same effect, or soak the circuit board in a saline solution,
or snip a wire on the capacitor that bypasses the series emitter
resistor in the audio driver, or just turn the knob maked "BASS"
counterclockwise. There's nothing about boron that has any special
effect on bass response. All that the boron does is change the gain and
bias of the transistors you inject it into. If you injected it into a
different set of transistors, you'd get different effects."
I don't even know that Melatonin Stimulating Hormone stimulates
melatonin production. To say "stimulates" asserts a direct cause-effect
relationship. Perhaps MSH raises the reference level for circulating
melatonin, or perhaps it lowers the sensitivity of some biochemical
receptor to the presence of melatonin, or perhaps it disturbs something
entirely unrelated to the system that produces melatonin, something to
which the melatonin-producing system responds by making more melatonin
as part of the process of counteracting the disturbance of some other
biochemical substance.
Without a schematic of the system, there's simply no way to decode any
apparent cause-effect relationship in a complex system. The apparent
relationships can be totally different from the actual relationships.
Consider this:
I pointed out that MSH might increase anxiety and by itself sustain
peeping. You reply was that MSH did not _increase_ peeping and that
therefore this effect could be ruled out. But that depends on your
model. You are imagining some direct cause-effect linkage, which leads
to predicting that an increase in the cause will produce an increase in
the effect, and a decrease will produce a decrease. But suppose that MSH
affects the gain of a control system. If the gain is high, the action of
the system will be appropximately equal and opposite to the effect of a
disturbance on the controlled quantity. Suppose the gain is 10. In that
case, the action will have about 90% of the effect of the disturbance,
leaving 10% error.
Now we do something that multiplies the gain of the control system by
10, from 10 to 100. If the system remains stable, we will now find that
the action cancels 99% of the effect of the disturbance. For the same
disturbance, the action has increased by 9% for a 1000% increase in
sensitivity. But suppose that our manipulation _reduces_ the gain by a
factor of 10, to 1. Now the control system will produce an action that
cancels only 50% of the effect of the disturbance, and the amount of
action will be only half the amount of the disturbance. The action has
been reduced by close to 50%. The error, which was less than 10 percent
of the reference signal, is now 50% of the reference signal, five times
as high.
Even though the system is made entirely of linear components, we get an
extremely nonlinear effect of changing the gain upward or downward. With
that kind of model in mind, what do we now expect if injecting MSH has
the effect of raising the gain of the control system?
I'm sure you can think of other nonlinearities that are to be expected
in a real neural system, such as the limit on maximum frequency of a
neural signal. My point is that as soon as you get away from a simple A-
causes-B mental model, the possibilities for explanations other than the
obvious one multiply, and it becomes evident that the simple empirical
evidence is of no use without the right model behind it.
Thousands of different brain operations employ similar chemistry. I
would guess that the main evolutionary reason for the existence of
different neurotransmitters is to reduce crosstalk between separated
systems, or, conversely, to provide information links between closely-
associated systems. But the particular brain chemicals, I strongly
suspect, have very little to do with the kind of function that is being
carried out. Brain chemicals are simply carriers of signals; the
particular chemical used as such a carrier is optional. So the fact the
MSH seems to be associated with a particular carrier is no more
significant than the fact that boron happens to be used to dope a
particular transistor to make it a semiconductor, or the fact that in
one transistor electrons carry the signal while in another, holes carry
the signal.
At the functional level (the level at which we observe behavior), the
nature of the signal-carriers is unimportant. To a biochemist, the exact
chemical nature of the processes is, of course, the main question, but
to a systems analyst it is irrelevant. Once we understand the system at
the functional level, we will be able to understand why one chemical or
another has the effects it has. But by the same token, it will be clear
that any number of other chemicals, or non-chemical phenomena, might
have performed the same functions.
Now, are you going to say, "yes, but it's still USEFUL to gather these
empirical facts"?
···
-----------------------------------------------------------------------
Best,
Bill P.