Some questions about reinforcment

[From Samuel Saunders (951212:16:57:17 EST)]

I promised not to be drawn in to discussion until I was ready to give a
reinforcement oriented critique and model for Rick Marken's proposed
experiment. With our recent Buffalo record snowfall, and the subsequent
disruption (such is children not in school yesterday or today and needing
supervision) I am afraid I am going to have to beg additional time anyway.
All that aside, I wanted to at least express an intent to respond to Bill
Power's's two questions. I will comment here on question 1, and reserve
question 2 for later comment.

Bill Powers (951210.2145 MST)
Every now and then, the apparatus delivers a pellet of food. It is said
that these deliveries are "contingent" on the pressing of the lever. I
ask, therefore, what you mean by "contingent," and how you would go
about proving that the deliveries are "contingent" on the contact
closures. There are two steps to this proof: first, defining
"contingent" so it has an unequivocal meaning, and second, proving (with
experimental evidence, not just logically or by appeal to common sense)
that the evidence fits the definition.

That B is contingent on A means occurrence of A is necessary for occurrence
of B. From the experimenters point of view, in the operant chamber,
contingency of food on switch closure is a matter of definition: the
apparatus is arranged so that it is _necessary_ for there to be a switch
closure before the food pellet is delivered. To prove this experimentally,
make a record of events occurring within the chamber. If food is contingent
on switch closure, then, looking backward in time in the record from any
occurrence of food, there will be an occurrence of a switch closure. In any
finite length record, one can potentially make an error if the rate of
switch closure is high relative to the rate of food delivery, as there may
never be recorded an instance of food delivery for which there is no
preceeding switch closure, even though switch closure may not be (always)
necessary for food delivery. If one can manipulate the rate of switch
closure, this potential error can be minimized, and practically eliminated
(if one records a segment substantially longer than any expected test
session with no switch closures, and never observers food delivery, then,
within the time frame of experimental interest, it should be safe to assume
that non-contingent food delivery does not occur).

In the operant chamber, where the experimenter controls the apparatus and
thus knows by definition, for each occurrence of food, whether that food
delivery was contingent on a switch closure, it is possible _for the
experimenter_ to define a 'degree of contingency' as the proportion of food
occurrences which were in fact contingent on switch closure.

The degree to which an organism exposed to the situation "contacts the
contingency" is a much more complicated question.

More to follow in a few days.

//Samuel Spence Saunders,Ph.D.