Arm-waiving

[From Bruce Abbott (960128.1400 EST)]

Bill Powers (960128.0530 MST) --

Bruce Abbott (960127.1525 EST)

    But Bill, this naive conception of the reinforcer is not the way it
    is generally viewed within EAB. What is missing is the whole
    process of evolution by which the organism was constructed.

"Was constructed?" How easy it is to pass from the world of modeling
into the world of theological arm-waving!

Wait a minute, let's back up here. We're looking for a set of reasonable
assumptions, not experimental proof that those assumptions are correct. The
question being addressed is "what is a reasonable way for organisms to be
constructed if they are to survive?" One can offer a logically consistant
answer without asserting that one has thereby discovered the correct answer.

So what do we know about organisms? Among other things, we know that to
survive they must take in substances, which we call nutrients, and avoid
taking in other substances, which we call toxins. We know that an organism
which failed to distinguish these two classes of substance would soon perish
from taking in toxins, should toxins be fairly prevalent in the organism's
environment. We know that organisms that die or are enfeebled by exposure
to toxins or by failure to ingest nutrients at an early stage in their lives
leave relatively few if any offspring, since life and health are
prerequisites for successful reproduction. Thus, we know that being able to
discriminate nutrients from toxins and to accept the former and reject the
latter are properties of organisms that would tend to promote their survival
as species.

If there is anything I have said so far that sound likes "theological
arm-waiving," I wish you would point it out to me, because all I hear are
established facts and cold logic.

To continue: We know that organisms have various sensory systems. In humans
these include four types of chemical receptor in the mouth that produce the
sensory qualities we label as sweet, sour, bitter, and salty. We know that
sweet receptors are activated by sugars, sour by acids, bitter by alkaloids,
and salty by salts. We know that in humans (and many other organisms),
sugars generally fall into the class of nutients and alkaloids (when present
in sufficient concentration) generally fall into the class of toxins. We
know that newborn human infants will readily accept sweet substances and
will reject bitter ones. They do it the first time these substances are
encountered.

These are just facts, not theological arm-waiving.

How is it that we humans come wired up to accept sweets and reject bitter
substances, right out of the box? One answer appeals to the process of
evolution as proposed by Charles Darwin. Another answer is that God made us
that way. Perhaps you have another answer. Whatever answer you (or I)
subscribe to, we can always point to the other person and accuse him or her
of "theological arm waiving." But so what? What is important is the set of
facts and the conclusions that can be deduced from them. The facts to which
my explanation appeals are true whichever theological position you hold, and
whichever arm you waive.

    To illustrate, certain physical events must occur if the organism
    is to survive, among them the periodic ingestion, in the right
    quantities, of appropriate nutrients. The word "appropriate" is
    important here because different organisms have different abilities
    to digest, absorb, and metabolize whatever ends up in their
    digestive systems. Organisms that ingest the right things tend to
    survive, those that don't, don't.

This is hardly an "illustration." Saying that "certain physical events"
must occur (and "appropriate" events at that) if the organism is to
survive is already a statement of faith in the purpose of evolution, as
if the goal is for some species to survive (I prefer saying "species,"
because organisms do not, typically, survive). What does "evolution"
(whoever he is) care if _any_ species survives?

How you got this out of my presentation is beyond me. If I say that you
must eat to survive, that is only a statement of fact. It is not "a
statement of faith in the purpose of evolution;" in fact it says nothing
about evolution at all. Go ahead, try it: don't eat for, say, 40 days and
see what happens!

Evolution is not a person, has no goals, and no cares. It is only the name
for a process. You put words in my mouth I never spoke and then criticize
me for saying them, and you wonder why you are so often accused of raising
and demolishing straw men. Well here's a fine example.

When you say "Organisms that ingest the right things tend to survive,
those that don't, don't," you are exposing the main weakness in naive
evolutionary theory. If eating were simply a chance outcome of random
behavior, it would not occur often enough for any species to survive.
There are far more non-nutritious objects than nutritious ones available
to eat, and they are far easier to catch than the nutritious ones. You
are saying only that species that have survived have survived. You have
not explained how it is possible that they have survived.

I am only stating the conditions necessary if they are to survive -- what
they must do. Although I stated it in the most general way possible (and
thus making the statement appear empty), it is easy to identify what
specific things a given organism requires if it is to survive.

You are correct that "if eating were simply a chance outcome of random
behavior, it would not occur often enough for any species to survive." In
fact, that is central to my thesis. Current organisms evolved from previous
versions, and those previous organisms already had the necessary mechanisms
or they would not have survived. Yet over the course of evolution, choice
of nutrients, ability to digest certain compounds and to tolerate others
that had been toxic to the organism's ancestors, ability to sense certain
correlated characteristics of nutrients and toxins, and so on did develop,
presumably via a process involving random variation and natural selection.
The details of this development are lost in history, but the result is here,
now. A consequence of this process of development is that current system
characteristics can be described in terms of function; one can ask, "what
does this characteristic, or this system, _do_ that enhances the organism's
reproductive fittness?" I have offered some speculative answers, but they
are logical and, I think, testable.

In order for either evolution or reinforcement to yield an organized
result, there must be a particular kind of object present, having a set
of critical properties already in place. The object must be able to
sense, to discriminate, and to act systematically by using stored or
ingested internal energy. Rocks do not evolve nor can they be reinforced
for behaving in any way but the way they already behave. They lack the
inner organization that is required to evolve or to be reinforced. Only
systems that are already alive can evolve or be reinforced. The whole
question is, "What is different about systems that can evolve and that
can be reinforced?" The answer that PCT proposes is that they are
control systems. But reinforcement theory has no answer. Nor does
evolutionary theory as it stands today.

Well, you have at least one thing right: there is a close parallel between
evolutionary theory and reinforcement theory. But you are wrong when you
state that neither evolutionary theory nor reinforcement theory have an
answer to your question. What is different about systems that can evolve?
They can vary, they can reproduce, and the variants can reproduce the
variations. What is different about systems that can be reinforced? They
can vary their actions, they can learn from the consequences of those
variations, and they can reproduce those variations. Of course, they must
have the right inner organization to do this; PCT offers one proposal.

All in all, I am disappointed that you chose to dismiss my proposal as
theological arm-waiving, especially given that it is, as I see it,
consistent with control theory. Mary (960128) seems to agree with me when
she says, in response to Chris Cherpas:

Emotional components of perceiving disturbances are part of the
process of physically mobilizing to do something about them, and
determining whether what one is doing is effective or not. What
better way to know something is wrong than to hurt?

So sensory inputs give rise not only to undifferentiated trains of neural
impulses that constitute the perception of the controlled variable, but to
evaluative signals consciously perceived as pleasure or pain? Now where
have I heard that before . . . .

Regards,

Bruce