Animistic Reinforcement

[From Rick Marken (941030.1345)]

Bruce Abbott (941029.1600 EST) --

I said:

And such an experimenter would object to being classed, along with
the forest people of New Guinea, as an "animist" (one who believes
that objects such as rocks and trees are alive and have souls) even
though this is precisely what he (or she) is -- a "Skinnerian animist".

To which Bruce apparently took offense . . .

Bruce says:

Rick, "offense" is too strong a term. "Mildly annoyed," perhaps, or
even "irritated," maybe, but not offended.

Why were you mildly annoyed or irritated by my comment? Were you also
mildly annoyed when Bill Powers (941029.1730 MDT) said:

This [Thorndike's] approach [to reinforcement] pays lip-service to
physical causality, but in a way that attributes to the physical
environment powers [ie. animism -- RM] that actually belong inside
of organisms.

which is exactly what I said about EAB researchers?

Bill Powers asked:

perhaps you can come up with a reinforcement simulation that will
produce the E. coli effect.

and Bruce Abbott (941029.1600 EST) answered:

The challenge is easily met. The program (in pseudo-code) is this:

If Going_up_gradient then continue;
If Going_down_gradient then tumble;
until Clock > EndSimulation

Bill Powers (941029.1730 MDT) replied to this, saying exactly what I
would have said (of course) which is:

What you have modeled is a control system; only a control system
can do what your stimulation-sketch describes.

The biggest problem posed by the E. coli effect for reinforcement
theory is neatly described by Bill as follows:

Only after the behavior has taken place (swim for some length of time
and then tumble) and the consequence occurs can the consequence be
known (end up swimming up or down the gradient, and thus
experiencing a positive or negative time rate of change of

Getting even more specific, Bill notes:

to produce a true reinforcement model, you have to show that a
_consequence_ of the current behavioral variables' setting modifies
the _next_ behavior, the next setting of that behavioral variable. The
only possible behavioral variable here is the delay prior to the next

So I look forward to seeing how the pseudo-code model you propose
can be shown to fit the definition of reinforcement.

By the way, have you actually tried doing the E. coli demo? If so, why
not try it again and, this time, while doing it, think about whether the
consequences (or antecedents for that matter) of your bar presses (the
direction of movement of the dot relative to a target position on the screen)
are selecting your bar pressing behavior or whether YOU are selecting the
consequences (the destination of the dot moving on the screen) of that
behavior. Does it seem like your behavior is selected BY its consequences or
does it seem like your behavior is better characterized as the selection OF
consequences? I know this is a very "subjective" approach to a question
that is best answered by modelling -- but I think you can get a FEEL for
the difference between the control theory and reinforcement theory views
of behavior if you try this.