Just say "no" to reinforcement

[From Rick Marken (970818.0930)]

Bill Powers (970818.0415 MDT) to Bruce Abbott --

Suppose the external observer describes the phenomenon this way:
"The environmental variable affected by the behavior is a
reinforcer, with the property that each time it is presented
when output A is connected (that is, when that contingency is
in effect), it increases the probability that output A will be
produced. Once output A has reached its steady state,
the reinforcer maintains the output at a constant level,
supporting output A."

Would you say that this description, including the implied
directions of causation, accurately portrays what the program
is actually doing?

Bruce Abbott (970818.1040 EST) --

No, for two reasons. First, there is a confusion about what
constitutes the reinforcer. You said earlier that the "program
that is organized to maintain one 'environmental' variable at a
reference level." Obviously this is the variable to be controlled.
The reinforcer would be whatever it is that an output produces,
which when produced, tends to drive the CV toward its reference
level. Second, knowing how you interpret the words "maintain"
and "support" in this context (as asserting direct causality; I
deny that this is necessarily true), I would answer "no, this
does not accurately protray what the program is doing," even if
the mistake leading to my first objection were corrected.

Boy, you are good.

Where do we go from here?

To just plain "no":wink:




Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net